Aeronautical

Telecommunications
Network
Advances, Challenges, and Modeling

Edited by
Sarhan M. Musa • Zhijun Wu
 Aeronautical
Telecommunications
Network
Advances, Challenges, and Modeling
 Aeronautical
Telecommunications
Network
Advances, Challenges, and Modeling

Edited by
Sarhan M. Musa
Prairie View A&M University, Houston, Texas, USA

Zhijun Wu
Civil Aviation University of China, Tianjin, China
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2016 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Version Date: 20150710

International Standard Book Number-13: 978-1-4987-0505-9 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-
ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may
rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or uti-
lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy-
ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the
publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923,
978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Dedicated to our families
Contents

Preface...............................................................................................................................................ix
Acknowledgments..........................................................................................................................xi
Editors............................................................................................................................................ xiii
Contributors....................................................................................................................................xv

1 Overview of Aeronautical Telecommunication Network............................................... 1

Sarhan M. Musa and Zhijun Wu

2 Optimization and Enhancement of MIPv6 in ATN....................................................... 13

Douzhe Li and Zhao Li

3 Modern Air Traffic Control Systems.................................................................................43

Yun-Fei Jia

4 Security of VHF Data Link in ATM..................................................................................65

Meng Yue

5 VDL2 Key Technology and Simulation............................................................................ 93

Gao Lin

6 GNSS Multipath Interference and Mitigation for UAVs in

Urban Canyon Environments........................................................................................... 137
Saeed M. Khan

7 Electromagnetic Interference to Aeronautical Telecommunications........................ 159

Zhigang Liu

8 Analysis and Modeling of the QoS Mechanism in ATN............................................ 191

Douzhe Li and Zhao Li

9 Time Division Multiplexing in Satellite Aeronautical

Communications System...................................................................................................219
Sarhan M. Musa and Zhijun Wu

10 ATN Transmission Control Algorithm Based

on Service-Oriented Architecture...................................................................................237
Haitao Zhang

vii
Preface

There are a lot of concerns today regarding the current transportation system growth,
procedures, and technologies in aeronautical telecommunications due to safety, flexibility,
and growing of new demands. Indeed, the recent growth in aeronautical telecommuni-
cation network (ATN) technology has changed the landscape and the role of different
systems that give aircraft a link with the ground while in flight. Increase in air traffic
transportation increases the communication rate of airborne aircraft for traffic coordina-
tion with the ground traffic controller.
It also gives the readers the opportunity to broaden and deepen their knowledge in the
advances, challenges, and modeling of ATN. Indeed, it covers recent and future methods
and problems illustrated in the field of ATN. It also provides strategies for integrating
existing and future data communications networks into a single internetwork serving the
aeronautical industry.
The book covers mobile Internet protocol version 6 (MIPv6) as a core of mobility man-
agement mechanism for aeronautical environment. The book deals with the simulated
performance of VDL Mode 2. It investigates the air traffic management (ATM) system
cyber-attacks since the 9/11 event and the current security situation of the ATM informa-
tion system related to the safety flight of air transportation.
The book provides an introduction to current trends in ATNs, including tools, tech-
niques, protocols, and architectures; MIPv6; air traffic control (ATC); security of ATM;
very-high-frequency (VHF) data link (VDL); aeronautical radio and satellite communica-
tions; electromagnetic interference to aeronautical telecommunications; quality of service
(QoS)-satisfied ATN routing mechanism speed dynamic environments; service-oriented
architecture (SOA)-based ATN transmission control algorithm; and the future directions,
opportunities, and challenges of ATN. This book has 10 chapters.
Chapter 1 provides an overview of ATN.
Chapter 2 presents an overview of MIPv6 and its extension that can be used in next-
generation ATN systems. The authors begin from the infrastructure of general mobile
communication, and then the motivation of MIPv6, MIPv6 protocol, and related triangu-
lar routing problem is presented. Due to some defects of the original MIPv6 protocol, four
types of extension are introduced, that is, Hierarchical Mobile Internet Protocol version 6
(HMIPv6), FMIPv6, PMIPv6, and network mobility (NEMO). Since NEMO is very suitable
for a whole mobile network moving, the authors focus on the optimization of NEMO and
its applications.
Chapter 3 introduces modern ATC systems from the perspectives of the user and the
developer. The composition of the ATC system is illustrated, and some peripheral equip-
ment such as automatic dependent surveillance–broadcast (ADS–B), secondary surveil-
lance radar, and surface movement radar are also provided. Some design experiences are
included in this chapter, including the considerations when selecting a better solution. All
subsystems of the ATC system are described in the order dataflow. In addition, the authors
show how to design an ATC system and how to use it to control in-flight airplanes.
Chapter 4 reviews the security of VHF data link in ATM. It analyzes some main security
threats for the Aircraft Communications Addressing and Reporting System (ACARS) data
link from the information perspective and offers a detailed solution to varieties of threats.

ix
x Preface

Meanwhile, it introduces varieties of algorithms used in the solution, such as encryption,

authentication, key exchange, and coding and decoding.
Chapter 5 expounds the application prospect of the VDL2 system and simulation. The
author provides the VDL technical advantage over the current ACARS and the necessity
to realize the system transition.
Chapter 6 presents the special case of Global Navigation Satellite System (GNSS) mul-
tipath interference in urban canyon environments starting with background information
and a discussion about the evolution of GPS and its interoperability with the Galileo sys-
tem under the conditions of continued modernization. Next, the author presents the results
of his own research to a spatial mitigation technique that capitalizes on the multiple GNSS
systems to cut down on the negative impact of multipath. It is expected that this treatment
will provide the readers a general understanding of the multipath phenomena and its
importance to unmanned systems.
Chapter 7 analyzes the electromagnetic interference of aviation communications equip-
ment, introduces the electromagnetic interference effects and response measures of avia-
tion communications equipment, and then analyzes the protective measures of aircraft
and ground stations against electromagnetic interference; it can be referenced for electro-
magnetic interference protection for aeronautical communications equipment.
Chapter 8 analyzes and models the QoS mechanism in ATN. The author addresses the
combination of the ATN environment with the IPv6/MIPv6/NEMO.
Chapter 9 provides the analysis of time division multiplexing (TDM) in a satellite aero-
nautical communications system when input sources are greater than available channels.
The analysis of blocking and clipping probabilities for TDMs was successfully achieved,
and results of the analysis were generated.
Finally, Chapter 10 introduces an ATN transmission control algorithm based on SOA.

Sarhan M. Musa
Zhijun Wu
Acknowledgments

Our sincere appreciation and gratitude to all the book’s contributors. We also acknowl-
edge the outstanding help and support of the team at Taylor & Francis Group/CRC Press
in preparing this book, especially from Nora Konopka, Michele Smith, Jonathan Plant,
Richard Tressider, Marc Gutierrez, Arlene Kopeloff, and Hayley Ruggieri.
Thanks to Jean Turgeon (JT), Liam Kiely, Marc Randall, Edwin Koehler, Paul Unbehagen,
Edwin Koehler, Mike Nelson, Brian Smith, Aaron Eddy, Roger Billings, Bryan Marklin,
Jeff Buddington, and Jeff Cox, Avaya, and Mike Bagby, Skynet Cloud Solutions, for their
advices and supports.
Thanks to Professor John Burghduff and Professor Mary Jane Ferguson for their support
and understanding and for being great friends. Dr. Musa thanks Dr. Mansoora A. Sheikh
for caring for his mother’s health during this project.
We appreciate the support received from Dr. Kendall T. Harris, dean of college of
engineering at Prairie View A&M University.
Finally, the book would never have seen the light of day if not for the constant support,
love, and patience of our families.

xi
Editors

Sarhan M. Musa, PhD, is currently an associate professor

in the Department of Engineering Technology, Roy G. Perry
College of Engineering, at Prairie View A&M University,
Texas. He has been director of Prairie View A&M Avaya
Networking Academy, Texas, since 2004. Dr. Musa has
published more than 100 papers in peer-reviewed jour-
nals and conferences. He is a frequently invited speaker,
has consulted for multiple organizations nationally and
internationally, and has written and edited several books.
Dr. Musa is a senior member of the Institute of Electrical and
Electronics Engineers (IEEE) and is also an LTD Sprint and a
Boeing Welliver fellow.

Zhijun Wu, PhD, currently serves as a professor in

the School of Electronics and Information Engineering
Technology at Civil Aviation University of China (CAUC),
Tianjin, China. He has been director of CNS/ATM Institute,
CAUC, since 2006. Professor Wu has published more than
50 papers in peer-reviewed journals and conferences. He is
a frequently invited speaker on electronics and information
technology, has consulted for multiple organizations nation-
ally and internationally, and has written and edited several
books. Professor Wu is a senior member of the Institute
of Communication, Tianjin, China, and a Boeing Welliver
fellow.

xiii
Contributors

Yun-Fei Jia Zhigang Liu

School of Electronics & Information Airport Management College
Engineering Technology Guangzhou Civil Aviation College
Civil Aviation University of China Guangzhou, People’s Republic of China
Tianjin, People’s Republic of China
Sarhan M. Musa
Saeed M. Khan Department of Engineering Technology
Department of Engineering Technology Prairie View A&M University
Kansas State University Salina Prairie View, Texas
Salina, Kansas
Zhijun Wu
Douzhe Li School of Electronics & Information
School of Electronic Information Engineering Engineering Technology
Tianjin University Civil Aviation University of China
Tianjin, People’s Republic of China Tianjin, People’s Republic of China

Meng Yue
Zhao Li School of Electronics & Information
School of Electronic and Information Engineering Technology
Engineering Civil Aviation University of China
Civil Aviation University of China Tianjin, People’s Republic of China
Tianjin, People’s Republic of China
Haitao Zhang
Gao Lin School of Electronic Information
School of Science and Technology Engineering
Tianjin Economic and Financial University Tianjin University
Tianjin, People’s Republic of China Tianjin, People’s Republic of China

xv
1
Overview of Aeronautical
Telecommunication Network

Sarhan M. Musa and Zhijun Wu

CONTENTS
1.1 Brief Introduction to Air Traffic Management................................................................... 1
1.2 CDM in Air Traffic Management.........................................................................................4
1.2.1 Definition and Introduction of CDM....................................................................... 4
1.2.2 Common Procedure of CDM....................................................................................5
1.3 Aeronautical Telecommunication Network........................................................................ 6
References........................................................................................................................................ 12

1.1  Brief Introduction to Air Traffic Management

In recent years, network-based, intelligent, automatic schemes are proposed to solve vari-
ous challenges and problems in the air traffic management (ATM) field [1,2]: this trend will
be sustained in the next decades or even in the future. For better understanding, a sum-
mary of ATM and its current development will be introduced in this chapter.
Before any description, a clear definition of ATM given by the International Civil
Aviation Organization’s (ICAO) document DOC4444 [3] should be pointed out, that is,
“the dynamic, integrated management of air traffic and airspace including air traffic ser-
vice, airspace management and air traffic flow management—safely, economically and
efficiently—through the provision of facilities and seamless service in collaboration with
all parties and involving airborne and ground-based function.” With a simple saying,
ATM can be treated as serials of systematic and complex operation on a huge system,
which manages countless aircraft and airlines.
ATM encompasses all systems and phases that assist every aircraft to depart from an
aerodrome, transit airspace, and land at a destination aerodrome, including air traffic con-
trol (ATC), aeronautical meteorology, air navigation systems (aids to navigation), airspace
management, air traffic services (ATSs), and air traffic flow management, or air traffic flow
and capacity management.
The increasing emphasis of modern ATM is on interoperable and harmonized systems
that allow an aircraft to operate with minimum performance change from one airspace
to another. ATC systems have traditionally been developed by individual states that
concentrated on their own requirements, creating different levels of service and capa-
bility around the world. Many air navigation service providers (ANSPs) do not provide
an ATC service that matches the capabilities of modern aircraft, so ICAO has developed

1
2 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

the Aviation System Block Upgrade initiative in order to harmonize global planning of
technology upgrades.
In the past decades, the currently used ATC system faces enormous challenges around
the world; the development of ATC lags behind the increasing traffic and forms a bottle-
neck. The most obvious is a wide range of flight delays that are mainly caused by complex
landing paths and waiting procedures; although safety is the primary consideration, the
effectiveness of flight is a disadvantage that cannot be ignored. Flight delay will lead to
huge economic losses and strong dissatisfaction of travelers. Excessive working pressure
of a controller also can cause serious security risks, since the approaching and landing are
almost dependent on the organization of the controller and are not automatic; when the
traffic flow is in a very high density, mistakes and emergencies are inevitable.
After entering the new century, the civil aviation world has realized the aforementioned
problems; also, many projects and researches have been carried on in recent years, such as
the following two famous projects:

1. Single European Sky ATM Research (SESAR). It is an ambitious initiative launched

by the European Commission in 2004 to reform the architecture of European ATM.
Its aim is using modern technology to construct a smart aviation traffic environ-
ment in the future.
2. The Next Generation Air Transportation System (NextGEN) is the name given to
a new national airspace system due for implementation across the United States in
stages between 2012 and 2025.

It can be inferred from SESAR and NextGEN that integration, interoperability, and seam-
lessness are the most important mechanisms in the further decades to build a modern
ATM system. Considering the bulky structure of ATM, in order to achieve these three
targets, many new technologies are used, including from physical radio frequency (using
higher frequency such as L band to avoid crowed in current VHF band), new wireless
data link (e.g., L band digital aeronautical communication system), MIPv6-based mobile
network, new surveillance technology (using automatic dependent surveillance [ADS]—
broadcast to replace secondary radar), and new computer systems, new management
application software, and so on.
The information acquired by various components and departments is integrated by effi-
ciently using a system-wide information management (SWIM) system. The core aim of
SWIM is sharing various types of information to support the implementation of collabora-
tive decision making (CDM). There are several important criteria to validate the ATM sys-
tem, such as real-time data sharing, running from terminal gate to terminal gate, 4D flight
path management, flexible and dynamic airspace allocation, and collaboration between
space and ground. Figure 1.1 gives a network-based ATM system structure.
As shown in Figure 1.1, there are four subsystems in ATM, that is, sky/ground integrated
ATS network, network precision air navigation system, system-wide multilevel monitor-
ing system, and collaborative working ATC system (please note that given the continuous
innovation, the name and duty of these subsystems could be changed). The connection
between system and people is all based on a network, which is the infrastructure of the
ATM system. In the past decades, aeronautical fixed telephone network was mainly used
by ATM to exchange various types of message, but there is a trend that the IPv6-based
Aeronautical Telecommunication Network (ATN/IPS) will be a promising technology in
the future.
For better understanding of Figure 1.1, a real world scenario is given in Figure 1.2.
Overview of Aeronautical Telecommunication Network 3

Connectivity
Connectivity between systems between people

Space/sky/ground
Network precision System wide Collaborative
integrated
air navigation multilevel worked ATC
air traffic service
system monitoring system system
network

Network-based air traffic management system (IPv6 based)

FIGURE 1.1
Framework of network-based air traffic management system.

Communications satellite

GPS
Satellite-based augmentation

Network precision air

navigation system

Space/sky/ground integrated
air traffic service network

System wide multilevel Cooperative airspace management

monitoring system

Surveillance technology based

on multipoint positioning
aerodrome
Area control center National traffic center
Aviation monitoring based on data
link and precise positioning Collaborative worked
ATC system

FIGURE 1.2
The typical running scenario of an air traffic management system.
4 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

As shown in Figure 1.2, numerous entities participate in the ATM system, such as space
relay/GPS satellites, aircrafts, terrestrial data link, and some management entities. All
time information exchanging is the critical point that can ensure the running of the whole
system. Another point that should be raised naturally is the decision process that is based
on the information from the ATM system, which is the CDM in Section 1.2.

1.2  CDM in Air Traffic Management

1.2.1  Definition and Introduction of CDM
The definition of CDM appears in ICAO’s document DOC 9971 [4], that is, “CDM is defined
as a process focused on how to decide on a course of action articulated between two or
more community members. Through this process, ATM community members share infor-
mation related to that decision and agree on and apply the decision-making approach and
principles.”
From the top-level viewpoint, the object of this collaborative process is to enhance the
holistic performance of the ATM system while balancing the needs of ATM community
individual members, for example, airport operators, aircraft operators, ground handlers,
and air traffic controllers.
CDM can be applied to all time frames of decisions, from long-range planning of sched-
ules to the tactical decisions of ground delay programs. Each user can participate to a
certain level that suits their operations and information requirements. However, in order
to maximize the benefits, it is important that all affected users participate in the informa-
tion sharing, and any member can propose a solution. The members that are commonly
included in CDM and their working flow are shown in Figure 1.3.

Users of
Airline Air traffic control sector Airport Pilot Controller Others
information

Collaborative ATM operational service

Integration of
ATC information processing
information

Space/sky/ground integrated
communication network

Infrastructure Various
Aeronautical Flight and flow Surveillance Meteorology
status information

Collaborative worked ATC system = Collaborative airspace management + collaborative traffic

management = High efficiency resource utilization + smooth running + safe flight

FIGURE 1.3
The collaborative work flow of collaborative decision making.
Overview of Aeronautical Telecommunication Network 5

As shown in Figure 1.3, we can see that CDM is a process of decision based on five
types of information: (1) aeronautical information, standards for aeronautical information
would be described through the aeronautical information exchange model (AIXM); (2)
flight and flow information, the evolution of Doc 9965 [5] would provide initial material to
define standards for the flight information exchange model; (3) surveillance information,
standards for the ground-to-ground exchange of surveillance information exist today; (4)
meteorological information, standards exist for the global dissemination of weather prod-
ucts; further standards development work may be required for new aviation weather prod-
ucts and to make these applicable to aviation CDM (e.g., see WXXM); and (5) infrastructure
status, standards for infrastructure status could largely be expressed using modified
AIXM standards.
Various data will be utilized by airlines, controllers, aircrafts, and so on. In one word, ratio-
nal decision should be based on rich and precise information. Some key technologies are
used in CDM to ensure this target, such as flexible use of airspace, approach and departure
route planning technology for terminal area or multiple airports, sector boundary division
technology, capacity assessment and airspace analysis technology, flight procedure design
based on multitype navigation, and theory and method for flight procedure assessment.

1.2.2  Common Procedure of CDM

In a collaborative process, the goal is achieving a desired outcome in the most efficient and
effective way for the organizations and for all collaborating parties involved. This can only
be achieved if the collaborating parties give as much attention to how they work together
throughout the process as they do to the process itself. Without one or the other, true coop-
eration, synergy, and teamwork, the goal can’t achieved.
To better understand CDM, some identification should be given: (1) actors (“Who is par-
ticipating in the collaboration?”); (2) roles and responsibilities (“What functions do the
actors perform and how do they interact?”); (3) Information requirements (description of
requirements and standards imposed on information exchanged as part of the aforemen-
tioned interactions); (4) making the decision (“How is a decision made?”); and (5) rules
(“What are some rules constraining the behavior?”).
The detailed steps can be summarized as six phases, that is, (1) need identification,
(2) analysis, (3) specification and verification, (4) performance case, (5) implementation
and validation, (6) operation, and finally subsequent and continuous maintenance and
improvement.

The Six Phases of CDM

The first phase is the identification of the need to apply CDM to realize a performance
improvement. This can relate to current processes/operations or to future processes.
A need statement should refer to the process(es) to which CDM should be applied and
need to specify the current situation, involving community members and current (or
projected) performance shortfall(s). It should include a first assessment (often based on
expert judgment) as to how and through which means CDM can mitigate the shortfall.
Shortfalls should be identified in areas related to all 11 key performance areas (KPAs)
identified in Doc 9854 [6]. While CDM has the ability to influence performance in all 11
KPAs, CDM provides a mechanism specifically well suited to addressing performance
areas, which are frequently difficult to quantify.
In the second phase, the CDM analysis, the target process is further analyzed from
a decision-making perspective. The analysis should make clear what decisions to be
6 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

made, which community members are involved (or affected), which information is used
in support of taking the decision(s), which process(es) is followed, how and through
which means the decision-making process can be improved, and how such an improve-
ment could contribute to a better performance.
The third phase, which builds on the CDM analysis, results in a shared and verified
specification of the CDM process. It will address among others

1. The decisions to be made and how they are reached and finalized
2. The community members involved and their roles/responsibilities in the
decision(s)
3. Agreement on objectives, where there may be a shared objective with indi-
vidual subobjectives (e.g., resolve congestion while minimizing impact to the
operation)
4. Decision-making rules, processes, and principles including specification of
timeline/milestones, interactions, roles, and responsibilities
5. Information requirements including data standards, quality, frequency, and
deadlines
6. CDM maintenance process: review, monitoring/verification, etc.

The objective of the performance case, developed through the fourth phase, is to jus-
tify the decision to implement the CDM process and to make the necessary invest-
ments. It should clearly specify what the costs are and the benefits (in the relevant
KPAs) that will result from the operation of CDM. It is important that the results of the
performance case are shared between all relevant community members. In the event
that the CDM process is an integral part of a new process, it should be integrated in
the performance case for that process.
The fifth phase, CDM validations and implementation, includes all steps to bring
CDM into operation. It includes training and informing staff and implementing/adapt-
ing systems and information networks.
Once the CDM process is operational (phase six), it should be subject to a continuous
and shared review and maintenance and improvement process. In this way, perfor-
mance can be continually improved.
Source: Adapted from ICAO, Manual on Collaborative Decision-Making (CDM), Doc 9971:
AN/485, 2012.

1.3  Aeronautical Telecommunication Network

As aforementioned, it is clear that space/sky/ground communication and information
sharing should abide by the principles effectively and safely; ATN that will be intro-
duced in this section is such a framework that can be used to ensure ATM and CDM
operations.
ATN is a network that is used by air traffic authorities. It is a global aviation standard
telecommunication network established by ICAO, and it is intended to provide seamless
air–ground and ground–ground. It uses digital data links to supplement voice communi-
cations and provide developed ATSs. The ATN manages the digital data transfer between
aircraft and the civil ATC facilities. The current infrastructure of ATN uses the Open
Systems Interconnection (OSI) Connectionless Network Protocol.
Overview of Aeronautical Telecommunication Network 7

ATN is a special aviation network that serves ATC units, aviation enterprises, and airline
companies. ATN is constituted by the following parts [1]:

1.
Air/Ground Communication Service Provider: It is responsible for providing the air–
ground access networks through terrestrial and satellite data link technologies.
2.
ANSP: It is responsible for managing the air traffic of one nation or region.
3.
Airline operation: It manages the commercial operation of an airline company.
4.
Mobile router: It is responsible for providing access to the air–ground data links for
the communication nodes of an airline.
5.
Home agent: It is a router or routing domain (RD) located at the home link of a
mobile node.
6.
ATS/Aeronautical operational control correspondent node: An ATS correspondent
node can be controlled to manage some specific airspace or other nodes to provide
information such as weather information. An Aeronautical Operational Control
Correspondent Node is located at an airline company or an airport to provide
services such as flight arrangement.
In an ATN route approach, the entire ATN is divided into many zones (ATN islands). In
each ATN island, there are several backbone RDs responsible for providing the default
routes to all the airplanes inside the island. A RD is the first layer default path provider in
ATN and is always able to achieve the paths reaching all airlines by some mechanisms.
The second default route provider is ATN Home. ATN Home is also a backbone RD that
provides the default paths reaching all airplanes of some airline companies. Indeed, ATN
Home can get the paths to all airlines of an airline company by some mechanism too [7–9].
ATN consists of a set of end systems (ESs) and a set of intermediate systems (ISs). ESs
are the source and destination of all data and are where the applications reside. ISs are
better known as routers and relay protocol data units from one system to another. The ESs
and ISs are organized into RDs. RDs are used to define sets of systems into clusters. These
clusters have two major properties: first, they are controlled by a single administration/
organization; second, a significant amount of the traffic is internal to the cluster [10].
Today, aeronautical communications use communications radio navigation and surveil-
lance bands and aeronautical radio spectrum as shown in Figures 1.4 and 1.5, respectively
[11], while Figure 1.6 shows the hypothetical ATN. The backbone router system is used in
designing the ATN to restrict the distribution of aircraft path information in order to avoid
the problems of flooding. However, it still allows every ground system the ability to send
data to any aircraft. There are four air–ground applications standardized by the interna-
tional standardization of ATN [12]:

1.
Digital flight information service: Allows pilots to request continual updates about
flight conditions (such as weather)
2.
Controller pilot data link communications (CPDLC): Allows replacing the present
controller/pilot voice interactions with digital messages to the extent possible
3.
Context management (CM): Allows supporting a dictionary service that lets the air-
craft log in whenever it enters a new ATS authority, exchange applications, and
associated network address information with the ground CM application server
4.
ADS: Allows position information determined by onboard navigation equipment
to be transferred to the ground.
8 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

90–110 130–535 kHz

NOB
30 kHz LORAN C 300 kHz

130–535 kHz 1,800–2,000 HF COMM

NOB
300 kHz LORAN A 3 MHz

3023 S&R 5680 S&R

HF COMM (2,850–22,000 kHz)
3 MHz 30 MHz

74.8–75.2 108 117.97 137

Marker Beacon VHF COMM
30 MHz VOR ILS 300 MHz

1,545– 1,559–
328.6–335.4 960 1,215 1,260 1,400 1,555 1,610 1,626.5 2,300–2,400 2,700–3,100
ILS Guide DME RNSS
slope
300 MHz 3 GHz

3,100 5,350–5,470 8,750–8,850 9,000–9,500 13,250–13,400

Surveillance radar MLS ASDE
3 GHz Radio altimeter 5,030–5,091–5,150 Airborne Airborne Precision 24.25– 30 GHz
weather Doppler approach 24.65
radar radar radar

Aeronautical communication systems

Aeronautical surveillance systems

Aeronautical radionavigation systems

FIGURE 1.4
(See color insert.) Communications radio navigation and surveillance bands.

ATN based on the Quality of Service (QoS) can be divided into four categories [9]:

1.
Residual error probability and transit delay: The routing decisions should favor low
residual error probability over low transit delay.
2.
Residual error probability and cost: The routing decision should favor low residual
error probability over low cost.
3.
Transmit delay and cost: The routing decisions should favor low transmit time over
low cost.
4.
Sequencing and transmit delay: The routing decisions should favor sending all
packets to the destination over a single path to maintain sequencing.
Overview of Aeronautical Telecommunication Network 9

Frequency
3 Hz 30 Hz 300 Hz 3 kHz 30 kHz 300 kHz 3 MHz 30 MHz 300 MHz 3 GHz 30 GHz

Band ELF SLF ULF VLF LF MF HF VHF UHF SHF

100,000 km 10,000 km 1,000 km 100 km 10 km 1 km 100 m 10 m 1m 10 cm 1 cm

Wavelength

117.975–137 MHz
AMS (R) Sat Comms
2.85–23.35 MHz VHF 1,544–1,555 MHz AMS Sat Comms
HF mobile 1,645.5–1,646.5 MHz 14–14.5 GHz

Mobile JTIDS/MIDS
Communications

SINCGARS HAVEQUIK

ILS glide beacon

VOR/ILS localizer beacon

MLS
DME
Radio navigation
NDB Rad Alt

LORAN C LORAN A
Marker beacon
GNSS
GNSS

Airborne weather Airborne weather

radar radar ASDR
SSR/ACAS 15.4–15.7 GHz

Surveillance
SSR/ACAS

Primary radar Precision Airborne weather

approach ASDR radar radar

FIGURE 1.5
(See color insert.) Aeronautical radio spectrum.
10 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Remote facility
edge base station Remote facility
edge base station
Hub

Remote facility Head

edge base station administrator Hub Remote facility
block edge base station

Core
network
possible
ATC control ATC disaster
Routing recovery site
options
triplication

Airport
Remote facility
edge base station Remote facility
edge base station

Hub Hub Hub

Remote facility
edge base station

Remote facility
edge base station Remote facility
edge base station

FIGURE 1.6
(See color insert.) Hypothetical Aeronautical Telecommunication Network. (From ICAO, Aeronautical Telecom­
munication Network (ATN): Manual for the ATN Using IPS Standards and Protocols, Doc 9896. September 2008.)

ATN is based on the Internet protocol suite (IPS) ATN/IPS [13]. The ATN/IPS protocol
architecture is illustrated in Figure 1.7. This architecture has four abstraction layers: the
link layer, the Internet layer, the transport layer, and the application layer.
As the ATN/IPS uses the 4-layer model of the Internet Engineering Task Force,
Figure 1.8 depicts the relationship between the OSI, ATN/International Organization for
Standardization, and the ATN/IPS protocols.
The ATN is enabler for the Communications Navigation and Surveillance (CNS)/ATM.
CNS/ATM integrates multiple communication and surveillance system such as GPS and
CPDLC.
There are several key technologies that should be further studied in ATN. They are
listed in Table 1.1. These technologies are separated by layer for better reading.
Overview of Aeronautical Telecommunication Network 11

Application Application
layer layer
OSI/PS OSI/PS
convergence convergence

Peer-to-peer connections

Transport layer Transport layer

(TCP, UDP) (TCP, UDP)
Inter-domain Inter-domain
router router

Internet layer Internet layer Internet layer Internet layer

(IPv6) (IPv6, BGP4+) (IPv6, BGP4+) (IPv6)

Link layer Link layer Link layer Link layer

Local or Local or
intradomain intradomain
subnetwork subnetwork
Interdomain
subnetwork

FIGURE 1.7
(See color insert.) Aeronautical Telecommunication Network/Internet protocol suite protocol architecture.
(From ICAO, Aeronautical Telecommunication Network (ATN): Manual for the ATN Using IPS Standards and Protocols,
Doc 9896, September 2008.)

OSI reference model ATN ISO protocols ATN IPS

ATN applications
Application layer
(e.g., CMIP, X.400) Consolidated
application
services ( e.g.,
Presentation layer Fast-byte SMTP, SNMPv3,
FTP, X.400 and
telnet)
Session layer Fast-byte

Transport layer
Transport layer TCP/UDP
Class 4

ES-IS/CLNP
Network layer SNDCF IPv6 & ICMPv6
X.25 PLP

Link layer LAP B (HDLC) LLC/MAC

Physical layer (e.g., Physical layer (e.g.,

Physical layer
ELA-232, V.35, X.21) FDDI, 802.X)

FIGURE 1.8
Protocol reference model. (From ICAO, Manual for the ATN Using IPS Standards and Protocols, ICAO, Montreal,
Quebec, Canada, 2011.)
12 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 1.1
Key Technologies in ATN
Name of Layer Key Technologies
Transport layer Service-oriented transmission control method in aerospace network environment
1. Performance analysis of cooperative aerospace information transmission
2. Qos routing mechanism based on a stable path
3. Service-oriented aerospace information transmission network control
Network layer QoS routing technology for high dynamic satellite network topology that includes
1. Network topology state prediction model
2. The estimate model for network link status information
3. Multiconstrained multipath routing algorithm based on the principle of
maximum probability
Link layer High-speed mobile aviation node broadband access technology that includes
1. Fast speed mobile node access control
2. Radio resource allocation to meet QoS requirements
3. Fast switching node based on the minimum cost

References
1. Cook, A. et al. Applying complexity science to air traffic management. Journal of Air Transport
Management 42, 49–158, 2014.
2. Di Gravio, G. et al. Overall safety performance of air traffic management system: Forecasting
and monitoring. Safety Science 72, 351–362, 2015.
3. International Civil Aviation Organization (ICAO). Procedures for Air Navigation Services: Air
Traffic Management, Doc 4444: ATM/501, 15th Edition, 2007.
4. International Civil Aviation Organization (ICAO). Manual on Collaborative Decision-Making
(CDM), Doc 9971: AN/485, 2012.
5. International Civil Aviation Organization (ICAO). Manual on Flight and Flow Information for a
Collaborative Environment (FF-ICE), Doc 9965: AN/483, 1st Edition, 2012.
6. International Civil Aviation Organization (ICAO), Global Air Traffic Management Operational
Concept, Doc 9854: AN/458, 1st Edition, 2005.
7. Qingbo, L., Xuejun, X., and Liang, Z. A mobility management mechanism in aeronautical tele-
communication network, IET International Communication Conference on Wireless Mobile and
Computing (CCWMC 2009), pp. 594–597, 2009.
8. Gallegos, S. K., Link, W. B., and Thomson, D. Innovation approaches for implementing the
aeronautical telecommunication network, 12th AIAA/IEEE Digital Avionics Systems Conference
(12th DASC), pp. 98–103, Fort Worth, TX, 1993.
9. Feighery, P., Hanson, T., Lehman, T., Mondrus, A., Scott, D., Signore, T., Smith, R., and Uhi, G.
The aeronautical telecommunications network (ATN) testbed. 15th AIAA/IEEE Digital Avionics
Systems Conference, Atlanta, GA, pp. 117–122, 1996.
10. International Civil Aviation Organization (ICAO), Aeronautical telecommunication network
(ATN) architecture plan for the ICAO Africa-India ocean (AFI) region, March 2012.
11. Stacey, D. Aeronautical Radio Communication Systems and Networks. John Wiley & Sons, Ltd.,
Hoboken, NJ, 2008.
12. Signore, T. L. and Girard, M. The aeronautical telecommunication network (ATN). IEEE Proceedings
of Military Communications Conference (MILCOM 98), vol. 1, pp. 40–44, Boston, MA, 1998.
13. International Civil Aviation Organization (ICAO). Aeronautical Telecommunication Network
(ATN): Manual for the ATN Using IPS Standards and Protocols, Part I, II, III. ICAO, Montreal,
Quebec, Canada, 2011.
2
Optimization and Enhancement of MIPv6 in ATN

Douzhe Li and Zhao Li

CONTENTS
2.1 Introduction........................................................................................................................... 13
2.2 Development of ATN Wireless Communication Standard............................................ 14
2.2.1 Framework of Aeronautical Mobile Communication Network......................... 15
2.2.2 MIPv6-Based Mobile Communication.................................................................. 18
2.3 Basics of Mobile IPv6............................................................................................................ 20
2.3.1 MIPv6 Protocol.......................................................................................................... 20
2.3.2 Understanding Details of MIPv6........................................................................... 21
2.3.3 Open-Source Simulation Tools............................................................................... 24
2.3.3.1 Using Direct Code Execution (DCE)....................................................... 25
2.4 Optimization of Triangular Routing in MIPv6................................................................ 26
2.4.1 Brief Introduction of Triangular Routing.............................................................. 26
2.4.2 Optimization of Triangular Routing..................................................................... 27
2.5 Enhancement of MIPv6........................................................................................................ 28
2.5.1 HMIPv6...................................................................................................................... 28
2.5.2 FMIPv6....................................................................................................................... 29
2.5.3 PMIPv6....................................................................................................................... 30
2.6 NEMO and NEMO Route Optimization Analysis.......................................................... 32
2.6.1 Introduction of NEMO and Its Relationship with MIPv6.................................. 32
2.6.2 Motivation and Scenarios of Using NEMO in Aeronautical Environment...... 35
2.6.3 NEMO Route Optimization and Its Application on ATN.................................. 36
2.7 Conclusion............................................................................................................................. 39
References........................................................................................................................................ 40

2.1 Introduction
The next-generation Aeronautical Telecommunication Network (ATN)1 is an all-IP
(Internet protocol) mobile communication network (abbreviated as ATN/IPS, here IPS
means Internal protocol suite). It will also conform to the concept of e-enabled aircraft
in the next 30 years.2 In the network layer, Mobile IPv6 (MIPv6) is selected as its mobile
communication/management protocol. MIPv6 allows mobile nodes (MNs, here refers to
aircraft stations) to remain reachable while flying around in the IPv6 network. Each MN is
always identified by its home address (HoA), regardless of its current point of attachment
to the network. While located away from its home, an MN will be associated with a care-

13
14 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

of address (CoA), which is configured by the attached network and provides information
about the current location of MNs.
In this chapter, first, some major concepts of general-purpose MIPv6 are described in
Sections 2.2 and 2.3. Then, we will combine the MIPv6 with its specific application in ATN,
especially elaborate some problems or enhancements of MIPv6 that should be focused on.
One major problem of MIPv6 is triangular routing, which depends on the relative loca-
tion of the home agent (HA); there can be a situation where the path in one direction (e.g.,
correspondent node [CN] to HA to MN) is significantly longer than the path in the reverse
direction (e.g., MN to CN). Detailed explanation and some solution schemes and relative
analytical model will be elaborated in Section 2.4.
Enhancements of MIPv6 such as FMIPv6, HMIPv6, and PMIPv6, which are used to
improve handover latency, reduce the amount of signaling required, and facilitate differ-
ent access technology, will be introduced and elaborated in Section 2.5, respectively.
Network mobility (NEMO) is also treated as an extension of MIPv6 and will allow any
node in the network that moves as a unit. It is very useful when MIPv6 is deployed in a
nautical environment. But an apparent drawback of NEMO is that it does not support
route optimization (RO), a feature that provides better end-to-end delay and path length
(in terms of number of hops); the optimization methods and related analysis will be given
in Section 2.6.

2.2  Development of ATN Wireless Communication Standard

There are two organizations that formulate mainly relevant standards, that is, the European
Organization for the safety of air navigation (EUROCONTROL) and the International Civil
Aviation Organization (ICAO).
EUROCONTROL’s Statistics and Forecast Service (STATFOR) has the goal of providing
statistics and forecasting on air traffic and monitoring the evolution of the air transport
industry. STATFOR provides short-term, medium-term, and long-term forecasts. Thereby,
different growth scenarios have been developed in order to account for different economic
progresses. The latest long-term forecast that was published in 20133 predicts air traffic
growth up to the year of 2050; according to the regulated growth scenario (i.e., growth
scenario C), it represents an extension to the existing environment we are in today. It is
characterized by moderate economic growth, with regulation reconciling the environmen-
tal, social, and economic demands to address the growing global sustainability concerns.
It exhibits a medium level of growth with 18.6 million Instrument Flight Rule movements
in Europe by 2050—2.0 times more than in 2012.
ICAO has recognized the need for technological progress. Specifically, this has been
discussed at the 11th Air Navigation Conference, where the conclusion was that the
aeronautical mobile communication infrastructure had to evolve in order to accommo-
date new functions of air traffic management (ATM) and to support air traffic growth.
As a result, EUROCONTROL and the U.S. Federal Aviation Administration decided to
establish a dedicated working arrangement (i.e., Action Plan 17,4 AP17). The objective
of that working group was the first step toward an enhanced aeronautical communica-
tion infrastructure, generically called Future Communication Infrastructure (FCI). The
FCI includes all components that are necessary for Air Navigation Service Provider
Optimization and Enhancement of MIPv6 in ATN 15

(ANSP), aeronautical operational centers, and aircraft to communicate with each other.
Thereby the so-called Future Radio System (FRS) is a subpart of the FCI. The FRS is
a term introduced by the Communications Operating Concept and Requirements for
the Future Radio System (COCR) document5 and indicates a globally consistent tech-
nological solution that provides appropriate communication infrastructure in order to
accommodate anticipated future air traffic growth. Thereby, the FRS is not necessarily
a single technology. AP17 has been a joint activity between Europe and the United
States, where several European and American partners worked in close cooperation in
order to assess and harmonize possibilities and solutions that would be appropriate for
the modernization of the aeronautical communication infrastructure. Thereby, it was
ascertained that the future ATM will face a paradigm shift from voice-based toward
data-based operations. AP17 was divided into several tasks, where six of them were
technically related. In the context of this chapter, the third task, investigation of new tech-
nologies for mobile communication is of special interest. This technical task investigated
newly emerging and existing data link technologies suitable to fulfill the requirements
of the FCI. Within this task, several technologies for different operational aspects were
compared.
A concurrent activity conducted by Working Group I of ICAO’s Aeronautical Commu­
nication Panel (abbreviated as ACP-WG-I) was the development of the ATN/IPS protocol
suite. Members of AP17 also concluded that the FCI will employ an ATN/IPS infrastruc-
ture. After the selection of the new communication systems has been made (which is still
not determined at the time of this book), the required Quality of Service (QoS) shall be
reassessed within an end-to-end IP environment. Since commercial-off-the-shelf (COTS)
products can facilitate development and ensure robustness, AP17 also recognized that it
would be beneficial if COTS products could be used or parts of COTS could be reused for
future aeronautical communication systems. It is crucial to realize the importance of that
statement. It means that COTS products shall be used wherever possible; however, if not
applicable, components of COTS products shall be reused if workable.
Some key recommendations were given in AP17: the development of an airport sur-
face communication system (i.e., AeroMACS),∙to finalize the selection of a terrestrial
data link communication system (i.e., L-DACS)6, and to consider satellite-based com-
munication systems, the prime candidate for oceanic and remote areas (c.f. ESA Iris).
Furthermore, AP17 concluded that it is important to retain QoS mechanisms provided
by any newly introduced data link technology when integrating the data link into a
subnetwork. This is imperative when considering the operation of IP over a specific
aeronautical link.

2.2.1  Framework of Aeronautical Mobile Communication Network

These data transmissions formed a network that contains wire and wireless communica-
tion. Cellular network has been already used in various wireless communication systems,
for example, personal mobile phone or police radio cluster communication. The cellular
concept can be applied to an aeronautical scenario in the same way. During the process
of a commercial flight, there are about five stages, namely, taxing, parking, takeoff, cruis-
ing, and landing. Transmitting data, voice, and other useful information between aircraft
and ground stations is almost dependent on wireless communication. So an aeronautical
mobile communication network, allowing transmission of bidirectional data via a termi-
nal device, need not be connected to a fixed physical link.
16 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Air–ground link Handover to

Hand-over satelite link

Polar or ocean area

ATC #2

ATC #1
ain
dom
Inter
ter
rou

Network domain #2
Network domain #1

FIGURE 2.1
(See color insert.) ATN mobile communication framework.

Controller–pilot voice communication is widely established throughout the aeronauti-

cal world and has had a very rapid increase in the number of subscribers to the various
ground station cellular networks over the last few years, although most of them are still
using analog HF band.
Aeronautical data communication has become a very important and rapidly evolving
technology as it allows users to transmit data from remote locations to other remote or
fixed locations. This proves to be the solution to the biggest problem of monitoring aircraft
status and supplying an Internet connection to traveling business people. Figure 2.1 gives
an illustration of the overall environment of ATN.
As we can see from Figure 2.1, different colors (orange/blue/green) indicate different
locations of an airplane. When the aircraft is flying over a continental area, the aircraft
station communicates with air traffic control (ATC) center #1 by using air-to-ground data
link, along with the path: airborne router → ground station → ground router (the orange
and blue routing path). Communication between planes can utilize air-to-air data link (air-
to-air link is out of scope of this chapter). In remote areas such as ocean and polar, you can
access ATN network via airborne router → satellite station → satellite data link → satellite
ground station (green routing path). In addition, each ground station can communicate
only with limited distance and a limited number of aircrafts, so handover between differ-
ent ground stations is necessary.
The aforementioned is a perceptual introduction. Here we give a further presentation of
the communication protocol suite of aeronautical mobile communication network shown
in Figure 2.2.
Optimization and Enhancement of MIPv6 in ATN 17

ATN applications ATN applications

Layer 5–7
OSI/IPS OSI/IPS
Application layer convergence convergence
Peer-to-peer
connections

Layer 4
Transport layer TCP, UDP TCP, UDP

Interdomain router Interdomain router

Layer 3
Network layer IPv6/MIPv6 IPv6, BGP4+ IPv6, BGP4+ IPv6/MIPv6

Layer 2 Wire link Wire link

Wire/wireless Wire/wireless
Link layer
link link

Layer 1
Local or Interdomain Local or
Physical layer
intradomain subnetwork intradomain
subnetwork subnetwork

FIGURE 2.2
Protocol architecture of ATN/IPS.

In Figure 2.2, the ground station, aircraft, satellites, or ATC centers (ATC#1, ATC#2), all
can be treated as IPS nodes, which have different responsibilities and functions. The access
and interconnection of ATN are realized by an IPS node, which is a device that implements
IPv6/MIPv6. There are two types of IPS nodes: (1) an IPS router, which forwards IP pack-
ets not explicitly addressed to itself, and (2) an IPS host, which can be treated as a message
source or a message destination, and it does not forward any packets.
We will now describe the internal structure of ATN/IPS from three viewpoints.
From the viewpoint of physical topology, the ATN/IPS internetwork is composed of IPS
nodes and a huge network which operate in a worldwide environment. The main function
of ATN is supporting air traffic service communication (ATSC) and aeronautical indus-
try service communication which contains aeronautical administrative communications
(AAC) and aeronautical operational communications (AOC).
From an administrative viewpoint, the ATN/IPS internetwork is composed of numerous
interconnected administrative domains. The concept of administrative domain is abstract,
but realistically, it may be an individual state, or a set of states that form an ICAO region,
an Air Communications Service Provider (ACSP), an ANSP, or any other organizational
entity that has the authority to manage ATN/IPS network resources and services. In each
administrative domain, there must exist several (also may be only one) IPS routers which
execute the interdomain routing protocol (such as the famous Border Gateway Protocol
version 4) specified in the ATN/IPS manual. Also, administrative domains and their coun-
terparts should coordinate their policy for carrying transit traffic.
From the viewpoint of routing, the routing information between autonomous systems
(ASs) can be exchanged by interdomain routing protocols. An AS is a connected group of
one or more IP address prefixes. The routing information exchanged includes IP address
18 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

prefixes of different lengths. For instance, the length of the IP address prefix exchanged
between ICAO regions may be shorter than the IP address prefix exchanged between indi-
vidual states within a particular region.
Mobility service providers (MSPs) will operate ATN/IPS mobility, which is based on
IPv6 mobility standards. An MSP in the ATN/IPS is an instance of an administrative
domain, which may be an ACSP, ANSP, airline, airport authority, government, or some
aviation organization. ATN/IPS MSP shall operate one or more HAs.

2.2.2  MIPv6-Based Mobile Communication

As we already know the mobility is braced by the IPS (i.e., IPv6 and MIPv6) protocol suite,
first, we give an intuitive metaphor. For better communication between nodes, each node
like the mailbox in front of our house should be identified with a mailing address and the
ZIP code. During mail delivery, the postman checks the address and then delivers mail to
the mailbox. Detailed information about IPS will be discussed as follows.
As proposed by the Internet Engineering Task Force (IETF) in RFC 2002 and subsequent
RFCs, mobile IP can provide an efficient, scalable mechanism for node mobility within the
Internet. Nodes may move and change their points of attachment to the Internet without
changing their IP address. This allows them to remain transparent to higher-layer connec-
tions while moving. Node mobility is realized without the need to propagate host-specific
routes throughout the Internet routing fabric. The MN uses two IP addresses: a fixed HoA
and a CoA that changes at each new point of attachment. This mechanism is briefly illus-
trated in Figure 2.3.
The current most implemented version of mobile IP is version 4 (MIPv4), which assumes
that a node’s IP address uniquely identifies the node’s point of attachment to the Internet.
A node must be located on the network indicated by its IP address in order to receive data-
grams destined for it; otherwise, datagrams destined to the node would be undeliverable.
If a node changes its point of attachment, in order not to lose its ability to communicate,
one of the two following mechanisms must typically be employed:

Air mobile node

Move Air mobile node

Home link Link B

Link A Router
Internet

Router
Link C
Router
Home agent

Correspondent node

FIGURE 2.3
Illustration of mobility and transparent data transmission.
Optimization and Enhancement of MIPv6 in ATN 19

Home network HA Remote network

CN
HoA

Foreign network

AR AR
CoA1 CoA2
MN

Movement

FIGURE 2.4
Simplified mobile IP.

1. The node must change its IP address whenever it changes its point of attachment.
2. Host-specific routes must be propagated throughout much of the Internet.

The first alternative makes it impossible for a node to maintain transport and higher-layer
connections when the node changes location. The second does not scale very well.
According to mobile IP protocol, an MN has two addresses: an HoA, which is a perma-
nent address, and a dynamic CoA, which changes as the MN changes its point of attach-
ment (See Figure 2.4; it is a simplified version of Figure 2.3). The fundamental technique
of mobile IP is forwarding. A CN, which is any peer node with which an MN is commu-
nicating, sends packets to the HA of the MN. The CN reaches the HA through normal
IP routing. Upon receipt of a packet from the CN, the HA will forward these packets to
the MN at its current CoA. The HA simply tunnels the original packet in another packet
with its own source address and a destination address of the current CoA. This is possible
because of the mobile IP protocol whereby the MN sends binding update (BU) messages to
the HA whenever its point of attachment changes. The BU associates the MN’s HoA with
its current CoA. The HA maintains a binding cache that stores the current CoA of the MN.
In order to describe the different functional building blocks of a typical aeronautical
wireless communication system, a generic functional architecture is used. Such a descrip-
tion consists of several entities that implement different functions. Thereby, entities may be
physically separated but may also be hosted by a single system. The various entities com-
municate among each other via interfaces (also known as reference points). These interfaces
are used to exchange user data (i.e., via the bearer plane) or signaling data (i.e., via the con-
trol plane). The relevant part for the operation of IP over an aeronautical link technology
is how the bearer plane from the mobile router (MR) to the access router (AR) is realized.
Figure 2.2 shows a generic functional architecture of a wireless communication system
with centralized medium access. In principle, each aeronautical communication system
could be mapped to such a generic functional architecture. The mobile station, which is
hosting an IPv6-capable router, is connected to the ground infrastructure of the commu-
nication system via the wireless link to the base station (BS). Within this context, the entity
that is controlling medium access to the physical transmission medium is generically
called the base station (BS). The wireless link itself is specifically designed for its purpose
20 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

and is specified through wireless interface specifications. Thereby, a link comprises the
physical layer and data link layer (DLL) definition. Mostly, such a standard includes a rudi-
mentary definition of an interface toward upper layers (i.e., the Internet layer), too.
A BS may communicate with other BSs through a specific interface. An inter-BS interface
may be interesting in order to speed up BS handovers, but it may also be interesting for a
DLL network. The BSs themselves are in general connected to a link-specific access gate-
way (LSAG) that is responsible for a domain as depicted in Figure 2.2. A domain within
this context is understood as a subnetwork of a larger network (i.e., the ATN). Thereby, the
LSAG may host other entities that are responsible for signaling, handovers, or resource
management. Furthermore, an AR is either directly attached to the LSAG or reachable via
an additional interface.
Theoretically, there may be a single BS responsible for a whole domain, especially if
satellite-based communication is considered. Ideally, a link technology design specifies a
service that fits exactly the needs of the upper layers. That is, the Internet layer uses the
specific link technology transparently, and all protocols implementing the service specifi-
cation of the Internet layer work properly. However, because network access providers, con-
tent providers, and customers have very different requirements for networks, it is hardly
possible to satisfy all needs accordingly. It might appear that assembling all requirements
for a safety-related aviation case would be simpler than in other commercially driven use
cases, as requirements should be clearly defined and available. However, due to the fact
that the aviation industry is evolving rather slowly (at least from a technological point of
view), it is hard to nail down all requirements for projects probably realized 20 years in
the future (if at all). Nevertheless, from the network point of view, one hard requirement
that should definitely be met by all future communication systems is full IPv6 compliance.
By strictly following this requirement, any future amendment of the ATN/IPS shall be
of no concern. Furthermore, it would be beneficial if IP multicast could be supported in an
efficient manner.

2.3  Basics of Mobile IPv6

2.3.1  MIPv6 Protocol
Before describing the working mechanism of mobile IPv6 (MIPv6) and its improvement,
some important terms in MIPv6 are adapted7 as follows.

• Mobile node (MN): A IPv6 node that is identified by its home address (HoA) dis-
regarding its current location. However, the MN should maintain reachability
using its HoA. When moving in the network, the MN may change its access links,
and then it will be associated with a new care-of address (CoA). By utilizing the
CoA, an MN has the awareness of its global address for the link that it is attached.
HoA and CoA is binded, and a tunnel is constructed between the HA and MN for
communication.
• Foreign link: A link that is in the MN’s visiting network, and is not the MN’s home
link.
• Care-of address (CoA): CoA is associated with the network that the mobile node is
visiting and provides information about the mobile node’s current location. The
entire datagram addressed to the MN’s HoA will be transparently forwarded to
Optimization and Enhancement of MIPv6 in ATN 21

the CoA. The CoA includes the foreign subnet prefix and an interface ID deter-
mined by the MN. This mechanism can realize the stateless address configuration.
MIPv6 allows the MN to be associated with multiple CoAs, but at the same time,
there is only one “primary” CoA, which is the one registered with the MN’s HA.
• Correspondent node (CN): An IPv6 node that communicates with an MN. A CN may
not have a mobile property, and may be a stationary PC. If the CN is an MN, it
should have the MIPv6 protocol stack installed and be away from home.
• Home link: The link that is affiliated to home subnet prefix. The HoA is within the
MN’s home link, also the HA always resides on the home link.
• Home address (HoA): The permanent address of the MN whenever the MN is in the
home network or in a foreign network; thus, HoA is always reachable. It can be
understood that the MN is always logically connected to the home link. An MN
can have multiple HoAs if there are several home subnet prefixes. When the MN is
attached to the home link, it will be operated like an IPv6 host, but when the MN
is away from home, the datagram destined to the MN should be forwarded from
the HA to the MN by a tunnel.
• Home agent (HA): A router resides on the MN’s home network. The HA records
the MN’s current location information (i.e, the CoA). When the MN is away from
home, it registers its current address with the HA, which forwards data sent to the
MN’s HoA to the MN’s current address on an IPv6 network through tunnels, and
also forwards tunneled data sent by the MN. A home agent may work in conjunc-
tion with a foreign agent, which is a router on the visited network.

Although the HA is a router connecting the home link to an IPv6 network, the HA does
not have to serve this function. The HA can also be a node on the home link that does not
perform any forwarding when the MN is at home.
MIPv6 allows an IPv6 node to be mobile to arbitrarily change its location on an IPv6 net-
work and still maintain existing connections. When an IPv6 node changes its location, it
might also change its link. When an IPv6 node changes its link, its IPv6 address might also
change in order to maintain connectivity. There are mechanisms to allow for the change
in addresses when moving to a different link, such as stateful and stateless address auto-
configuration for IPv6. However, when the address changes, the existing connections of
the MN that is using the address assigned from the previously connected link cannot be
maintained and are ungracefully terminated.
The key benefit of MIPv6 is that even though the MN changes locations and addresses,
the existing connections through which the MN is communicating are maintained. To
accomplish this, connections to MNs are made with a specific address that is always
assigned to the MN and through which the MN is always reachable. MIPv6 provides
transport layer connection survivability when a node moves from one link to another by
performing address maintenance for MNs at the Internet layer.

2.3.2  Understanding Details of MIPv6

In this section, we use the knowledge gained from previous sections to draw a simplified
flowchart for the MN’s and CN’s operation. This flowchart can be compared to the one
used in Section 2.2 for normal IPv6 hosts. Since we want to focus the mobile signaling
mechanism, the BU security details were not considered in this section. Our flowchart for
the MN’s sending implementation is shown in Figure 2.5.
22 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Receive upper layer data

Construct IP header

BUL entry exists?

Yes No
Find CoA

Put CoA in HAO

BC entry exists?

Yes No
Put CoA in RH
and HoA in dst

IP sec needed?
Yes
Add
No AH/ES

HAO exists?

No
Swap HAO content
with src address

RH type 2 exists?

Yes No
Swap RH content
with src address

Detemine next hop

Send packet
(a)

FIGURE 2.5
(a) Processing sending packets at the correspondent node.  (Continued )
Optimization and Enhancement of MIPv6 in ATN 23

Receive packet from link layer

Process IP header

RH type 2 exists? BC: Binding cache

CoA: Care-of address
Yes HAO: Home address option
Swap HoA with dst No HoA: Home address
address RH: Routing header

HAO exists?
BC entry
exists?

Yes
Swap HoA and No
src address

Send binding error

Ipsec needed?

Yes
Verify
No AH/ESP

Pass to upper layers

(b)

FIGURE 2.5 (Continued)

(b) Processing receiving packets at the correspondent node.

When the MN’s IP layer implementation receives upper-layer data, it will have already
supplied the HoA as a source address for applications. Hence, an IP header can be con-
structed with the HoA as the source address and the CN’s address as the destination
address. Next, the BU list is checked to see if the MN has already sent a BU to the CN and
get the CoA from the correspondent’s node’s entry. If it has, the MN can include its HoA in
the HoA option. However, for now, the MN places its CoA in the HoA option instead. We
see why this is done when we analyze the following steps in the flowchart. The MN then
checks its binding cache to see if the CN has sent it a BU. Note that CN refers to a func-
tional entity in an IPv6 node and does not exclude a CN from being an MN as well. If the
CN had sent a BU to the MN, a binding cache entry is found. Hence, the MN constructs a
routing header type 2 and places the CN’s CoA inside it.
Assuming that the MN does not need to add any other extension headers, the next
step is to check whether IPsec (IP Security, a protocol suite for securing IP communica-
tions, details not covered in this chapter) is needed. This is done by checking the security
policy database to see whether the packet should be protected. When IPsec is used, the
destination options header containing the HoA option needs to be placed before the
IPsec header (AH or ESP). This is done because IPsec identifies a security association
in the security association database by selectors. Selectors include the source and des-
tination addresses of the nodes sharing the security association. To avoid changing the
security association every time the MN moves, it is best to use the HoA (not the CoA) as
a selector identifying the MN. When the packet is received by the CN, the headers are
24 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

processed in the order they appear in the packet; therefore, IPsec headers are processed
last, before the upper layer. Hence, an IPsec implementation in the CN always sees the
HoA in the source address field of a received packet. Now we can see the reason for
keeping the MN’s HoA in the source address field and its CoA in the HoA option until
IPsec has added its header. We need to perform the IPsec operations on the header seen
by the CN after verifying the IPsec header. That is, the packet needs to look exactly the
same when passed to the CN’s IPsec implementation as it did when passed to the MN’s
IPsec implementation.
Otherwise, to IPsec, the packet will seem to have been modified and will be dropped.
After IPsec processing is done, the MN swaps the contents of the HoA option and the
routing header with the source and destination addresses respectively. The next hop (the
next destination of the data packet) is determined based on whether the packet is tunneled
to the HA or sent directly to the CN, then the addresses of next hop are tied to interfaces.
For example, when the MN is on a foreign link, its HoA is associated with a tunnel interface
to the HA. When it moves back home, the tunnel is deleted, and the HoA is associated with
a physical interface. The tunnel interface can be treated by the IP layer implementation as
another physical interface that is one hop away from the HA. Hence, if the packet contains
the HoA in the source address field, it is immediately tunneled to the HA; otherwise, it is
sent on the interface associated with the CoA.
Now let’s consider the steps taken by the CN when it receives packets from the MN, as
shown in Figure 2.5b.
When the CN receives a packet, it processes the headers in the order that they were
sent. If a routing header type 2 is included, and provided that there is one address in
it and that it is the CN’s HoA, it swaps that address with the destination address in the
packet; otherwise, the packet is dropped. The node then processes other extension head-
ers. In Figure 2.5, we skip this processing until it gets to the destination options header. If
an HoA option is found, the binding cache is searched for a corresponding entry. If noth-
ing is found, a binding error message is sent to the source address in the IPv6 header. If
the CN does not implement any MIPv6 messages, then it does not understand the HoA
option and sends an ICMP error message to the source address in the packet. If a bind-
ing cache entry is found for this HoA, the home address is swapped with the CoA in the
source address field.
If an IPsec header is included, it is verified as usual. Note that the packet would now look
exactly the same when seen by IPsec implementations at both ends. Finally, the informa-
tion is passed to upper layers.

2.3.3  Open-Source Simulation Tools

Open-source network simulator tools are often used in academic researching; although
open-source software does not have the privilege of customer service, it’s easy for mod-
ification, realizing, and testing a researcher’s novel idea; also it can help the researcher
to understand the whole network framework. Usually there are two famous open-
source network simulators mentioned in academic papers, that is, OMNET++ and NS
(NS-2 or NS-3).
The NS-3 simulator is a discrete-event network simulator tool, which is typically used
for research and educational purpose. The NS-3 project is an open-source project, which
was started in 2006, and the current version is NS-3.21 (until Sep. 2014). Some tips as fol-
lows can help us understand NS-3.
Optimization and Enhancement of MIPv6 in ATN 25

• NS-3 is not an extension of NS-2. It is a new simulator. Both the simulators are
written in C++, but NS-3 is a new simulator that does not support the NS-2 APIs.
Some models from NS-2 have already been ported from NS-2 to NS-3. The project
will continue to maintain NS-2 while NS-3 is being built, and will study transition
and integration mechanisms.
• NS-3 is open source, and the project strives to maintain an open environment for
researchers to contribute and share their software.
• In NS-3, the simulation code should be written in C++ (recommended) or Python;
NS-3 dropped the TCL simulation code, which was widely used in NS-2. So profi-
cient C++ programming skills can help to quickly grasp NS-3.

For the purpose of simulating ATN/IPS, there is nothing difficult in downloading and
compiling NS-3 for preparation. The importance is currently there is not an official MIPv6
simulation environment in the latest release. MIPv6 and NEMO simulations are base on a
third-party component called Usagi-Patched MIPv6 stack (UMIP), which was written by
a Japanese programmer.

2.3.3.1  Using Direct Code Execution (DCE)

The DCE NS-3 module provides facilities to execute existing implementations of userspace
and kernel space network protocols within NS-3.
The motivation and advantages of utilizing DCE are

• If you do not want to maintain multiple implementations of a single protocol or

re-implement complicated protocols such as Open Shortest Path First (OPSF) or
Border Gateway Protocol (BGP) routing
• If you want to debug and test your code within a controlled environment
• If you want to create a miniature network (let’s say, a model of an ISP network) in
a single node

As of today, the Quagga routing protocol implementation, the CCNx CCN implementa-
tion, and recent versions of the Linux kernel network stack are known to run within DCE,
hence allowing network protocol experimenters and researchers to use the unmodified
implementation of their protocols for real-world deployments and simulations.
Using UMIP: UMIP is an open-source implementation of MIPv6 and NEMO Basic
Support for Linux. It is released under the GPLv2 license. (Please visit http://umip.org/
for more recent news and updates; latest update is UMIP 1.0.) The UMIP support on DCE
enables the users to reuse routing protocol implementations of UMIP. UMIP now supports
MIPv6 (RFC3775), NEMO (RFC3963), Proxy Mobile IPv6 (PMIPv6; RFC5213), RFC3776 and
RFC4877 (IPsec and Internet Key Exchange version 2nd [IKEv2]), etc., and these proto-
col implementations can be used as models of network simulation. It reduces the time of
reimplementation of the model and potentially improves the result of the simulation since
it is already actively running in the real world.
Unlike the Quagga support of DCE, the UMIP support of DCE requires Linux kernel
integration of DCE, because the dependencies of UMIP, UMIP must interact with Linux
kernel.
26 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

2.4  Optimization of Triangular Routing in MIPv6

2.4.1  Brief Introduction of Triangular Routing
While running mobile IP protocol, an MN owns two addresses: an HoA, which is a perma-
nent address, and a dynamic CoA, which changes as the MN changes its point of attach-
ment. The fundamental technique of mobile IP is forwarding. A CN, which is any peer
node with which an MN is communicating, sends packets to the HA of the MN. The CN
reaches the HA through normal IP routing. Upon receipt of a packet from the CN, the HA
will forward these packets to the MN at its current CoA. The HA simply encapsulates the
original packet in another packet with its own source address and a destination address
of the current CoA. This scheme is realized by BU messages, which are sent by MN to the
HA whenever MN’s point of attachment changes. The BU associates the MN’s HoA with its
current CoA. The HA maintains a BC that stores the current CoA of the MN. In the reverse
direction, the MN could simply send packets directly to the CN using normal IP routing.
While all traffic from the CN to the MN is routed via the HA, then this situation is
called triangular routing. This type of routing increases the traffic on the network as the
packets are first routed to the HA, and from here, they are transmitted to the MN by a tun-
nel. In particular, the load on the HA will be increased significantly. As we can see from
Figure 2.6, the triangular routing problem is demonstrated by a very far distance between
the MN (i.e., aircraft node) in America and the HA in Asia; all the data from MN to CN
should be forwarded by HA. This process will consume much network resource. While
with optimization, the data can be directly forwarded to CN.

China America

CN, Correspondent node

Destination IP host in
session with a mobile
node

Without
optimization
With
optimization

HA, home agent

Maintains an
association between
the MN’s “home” IP address MN, Mobile node in the air
and its care of address on the An IP host that
ATC center maintains network connectivity using its
foreign network
in China “home” IP address, regardless of which
link (of network) it is connected to

FIGURE 2.6
Solution of triangular routing.
Optimization and Enhancement of MIPv6 in ATN 27

However, this results in triangular routing, and depending on the relative location of the
HA, there can be a situation where the path in one direction (e.g., CN to HA to MN) is sig-
nificantly longer than the path in the reverse direction (e.g., MN to CA). A further consid-
eration in this case occurs if the MN uses its HoA as the source address. The problem here
is that many networks perform ingress filtering of incoming packets and will not accept
packets that are topologically incorrect. This would be the case with packets from the MN
because they actually originate from the CoA, but the source address in the IP packet is
the HoA.

2.4.2  Optimization of Triangular Routing

In the reverse direction, the MN could simply send packets directly to the CN using nor-
mal IP routing. However, this results in triangular routing, and depending on the relative
location of the HA, there can be a situation where the path in one direction (e.g., CN to HA
to MN) is significantly longer than the path in the reverse direction (e.g., MN to CN). A
further consideration in this case occurs if the MN uses its HoA as the source address. The
problem here is that many networks perform ingress filtering of incoming packets and will
not accept packets that are topologically incorrect. This would be the case with packets
from the MN because they actually originate from the CoA, but the source address in the
IP packet is the HoA. Because of these issues, MIPv6 allows the MN to follow the same
path back to the CN via the HA using bidirectional tunneling whereby in addition to the
HA tunneling packets to the MN, the MN reverse-tunnels packets to the HA. The HA will
decapsulate a tunneled IP packet and forward it to the CN. With bidirectional tunneling,
the CN is not required to support mobile IP.
Bidirectional tunneling solves the problems of triangular routing and ingress filtering;
however, there still can be suboptimal routing since the path from the MN to the CN via
the HA may be relatively long even when the MN and the CN are in close proximity.
With MIPv6, the situation where the path through the HA is longer than a direct path to
the CN may be addressed using a technique called route optimization. With RO, the MN
sends BUs to both the HA and the CN. In this case, the MN and the CN can communicate
directly and adapt to changes in the MN’s CoA. RFC 3775 defines the procedures for RO.
It requires that the MN initiate the return routability procedure. This procedure provides
the CN with some reasonable assurance that the MN is addressable at its claimed CoA and
its HoA.
It is generally acknowledged that there are drawbacks to RO. RFC 4651 presents a
taxonomy and analysis of enhancements to MIPv6 RO. This document notes that the
two reachability tests of the return routability procedure can lead to a handoff delay
unacceptable for many real-time or interactive applications, that the security and the
return routability procedure guarantees might not be sufficient for security-sensitive
applications, and that periodically refreshing a registration at a CN implies a hidden
signaling overhead. Because of the overhead and delay associated with the return
routability procedure and because at least for ATSC it is expected that the CN and the
HN will be in relative close proximity, this manual requires that IPS CNs that imple-
ment MIPv6 RO allow RO to be administratively enabled or disabled with the default
being disabled. New solutions to RO are expected as a result of IETF-chartered work
in the mobility extensions for IPv6 working group, which includes aviation-specific
requirements.
28 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

2.5  Enhancement of MIPv6

2.5.1 HMIPv6
One technology for these issues is hierarchical MIPv6 (HMIPv6) (RFC 5380). RFC 5380 intro-
duces extensions to MIPv6 and IPv6 neighbor discovery to allow for local mobility han-
dling. HMIPv6 reduces the amount of signaling between an MN, its CNs, and its HA.
HMIPv6 introduces the concept: the mobility anchor point (MAP). A MAP is essentially a
local HA for an MN.
When an MN visits a new access network and enters a MAP domain, it will receive
router advertisements that include information about one or more local MAPs. The MN
can bind its current location, that is, its on-link care-of address (LCoA), with an address on
the MAP’s subnet, called a regional care-of address (RCoA). Actually LCoA is a CoA used
in the binding update to MAP, while RCoA is used for route optimization, and it is a CoA in
the binding update to home agent and corresponding nodes. MAP can be treated as a local
HA, and it will receive all packets in order to the MN it is serving and will encapsulate and
forward them directly to the MN’s current address. If the MN does not change its current
address within a local MAP domain (LCoA), it needs not to register only the new address
with the MAP. The RCoA will change only if the MN moves within a MAP domain.
RFC 4140 notes that the use of the MAP does not assume that a permanent HA is present,
that is, an MN need not have a permanent HoA or HA in order to be HMIPv6-aware or use
the features of HMIPv6. HMIPv6-aware MNs can use their RCoA as the source address
without using an HoA option. In this way, the RCoA can be used as an identifier address
for upper layers. Using this feature, the MN will be seen by the CN as a fixed node while
moving within a MAP domain. This usage of the RCoA does not have the cost of MIPv6
(i.e., no bindings or HoA options are sent back to the HA), but still provides local mobility
management to the MNs with near-optimal routing although such use of RCoA does not
provide global mobility.
Here we will bring in the concept of a hierarchical MIPv6 network. A new node is intro-
duced named as the mobility anchor point (MAP).
An MIPv6 node can move within the Internet topology while maintaining reachability
and ongoing sessions between the MN and CN according to MIPv6 (RFC3775). To fulfill
this, every time an MN moves to another point, it should send BUs to its HA.
The time delay often happens in the process of signaling exchange. For example, after
sending the BU, the MN may send datagrams by means of its HA immediately, but the
HA cannot send any datagrams back to the MN before it receives the BU. This will cause
at least half a round-trip delay before packets reach the right place again. We will have
another delay event for sending data packets if the MN chooses to wait for a binding
acknowledgment (BA) returned. The worst situation always occurs when the MN and its
HA are in different and distantly located across the world, so the round-trip times will be
more longer.
Additional delay will also occur if the MN adopts a route optimization mechanism.
Authenticating BUs requires about 1.5 round-trip times between the MN and each CN.
(Assume in the ideal environment, the entire return routability procedure has no packet
loss.) We can do it in parallel by sending BUs to the HA, and there will be further optimiza-
tions that reduce the required 1.5 round-trips (RFC4449, RFC4651, and RFC4866).
Nevertheless, the signaling exchanges need to update the location and they will always
cause some disruption to active connections. Some packets will be lost. There may be
Optimization and Enhancement of MIPv6 in ATN 29

effects to upper-layer protocols along with link layer and IP layer connection setup delays.
We will enhance the performance of MIPv6 if we reduce these delays during the time-
critical handover period.
Moreover, as for wireless links, such a solution reduces the number of messages which
are sent over the air interface to all CNs and the HA. A local anchor point will also allow
MIPv6 to benefit from reduced mobility signaling with external networks.
For these reasons, a new MIPv6 node with the name of the mobility anchor point is used
and can be located at any level in a hierarchical network of routers, including the AR. The
MAP will limit the amount of MIPv6 signaling outside the local domain.
The introduction of the MAP provides a solution to the issues outlined earlier, in the
following way:

• The MN sends BUs to the local MAP instead of the HA (which is typically further
away) and CNs.
• We should transmit the only BU message by the MN before traffic from the HA
and all CNs is rerouted to its new location. This is independent of the number of
CNs with which the MN is communicating.

Substantially, A MAP is a local HA. If we want to improve the performance of MIPv6

while minimizing the impact on MIPv6 or other IPv6 protocols we should introduce the
hierarchical mobility management model in MIPv6. For some security considerations,
while using MIPv6 RO, HMIPv6 allows MNs to hide their location from CNs and HAs.
We should notice that the use of the MAP does not depend on, or assume the presence
of, a permanent HA. On the other hand, an MN need not have a permanent HoA or HA
in order to be HMIPv6-aware or use the features in this specification. An MN may use
a MAP in a nomadic manner to acquire mobility management within a local domain.
Section 6.5 describes such a scenario.

2.5.2 FMIPv6
The combined handover latency is often sufficient to affect real-time applications, which
is suitable especially in aeronautical environment. Throughput-sensitive applications
can also benefit from reducing this latency. Fast Handovers for MIPv6 (FMIPv6) [RFC
4068] describes a protocol to reduce the handover latency. It is a further enhancement
to MIPv6.
FMIPv6 attempts to reduce the chance of packet loss through low-latency handoffs. It
attempts to optimize handovers by obtaining information for a new AR before disconnect-
ing from the previous AR. ARs request information from other ARs to acquire neighbor-
hood information that will facilitate handover. Once the new AR is selected, a tunnel is
established between the old and new routers. The previous care-of address is bound to a
new care-of address so that data may be tunneled from the previous AR to the new AR
during handover. Combining HMIPv6 and FMIPv6 would contribute to improved MIPv6
performance, but this comes at the cost of increased complexity.
Two requirements are considered when developing an FMIPv6 protocol. One is how we
can permit an MN to send packets when it detects a new subnet link. The other is how we
can deliver packets to an MN when its attachment is detected by the new AR.
The protocol defines IP protocol messages necessary for its operation in spite of link
technology. It does not depend on specific link-layer features when it allows link-specific
30 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

customizations. By definition, it employs handovers that interwork with mobile IP. Relying
on its new AR, an MN engages in mobile IP operations including return routability
(RFC3775). There are no special requirements for an MN to behave differently with respect
to its standard mobile IP operations.
FMIPv6 is applicable when an MN has to perform IP-layer operations as a result of han-
dovers, but it does not address improving the link-switching latency. It does not modify or
optimize procedures related to signaling with the HA of an MN. Indeed, while targeted
for MIPv6, it could be used with any mechanism that allows communication to continue
despite movements.

2.5.3 PMIPv6
In MIPv6, as described earlier, the MN updates the HA with BU messages. This mode of
operation is called node-based mobility management. A complimentary approach is for
access network service providers to provide network-based mobility management using
PMIPv6 on access links that support or emulate a point-to-point delivery. Except from the
simulation of MIPv6 by the aforementioned tool NS-3 (see Section 2.3.3), Hyon-Young Choi
has also given a report on how to program an NS-3-based PMIPv6 simulation environment.8
He also created a site for sharing his codes (https://sites.google.com/site/pmip6ns/).
MIPv6’s approach that supports mobility does not require the MN to be involved in the
exchange of signaling messages between itself and the HA to potentially optimize the access
network service provision. A proxy mobility agent in the network performs the signaling
with the HA and does the mobility management on behalf of the MN attached to the net-
work. The core functional entities for PMIPv6 are the local mobility anchor (LMA) and the
mobile access gateway (MAG). The LMA is responsible for maintaining the MN’s reach-
ability state and is the topological anchor point for the MN’s home network prefix(es).
The MAG is the entity that performs the mobility management on behalf of an MN, and
it resides on the access link where the MN is anchored. The MAG is responsible for detect-
ing the MN’s movements to and from the access link and for initiating binding registra-
tions to the MN’s LMA. An access network that supports network-based mobility would
be agnostic to the capability in the IPv6 stack of the nodes that it serves. IP mobility for
nodes that have mobile IP client functionality in the IPv6 stack as well as those nodes that
do not would be supported by enabling PMIPv6 protocol functionality in the network.
The advantages of PMIPv6 are reuse of HA functionality and the messages/format used
in mobility signaling, and common HA would serve as the mobility agent for all types of
IPv6 nodes. PMIPv6 like HMIPv6 is a local mobility management approach that further
reduces the air-ground signaling overhead.
The detailed information of IP mobility for IPv6 hosts is described in MIPv6 (RFC3775).
MIPv6 needs client functionality in the IPv6 stack of an MN. We can interchange signaling
messages between the MN and HA to enable the creation and maintenance of a binding
between the MN’s HoA and its CoA. In order to realize mobility, according to RFC3775, the
IP host needs to transmit IP mobility management signaling messages to the HA, which
is located in the network.
There is another approach to solving the IP mobility challenge, namely network-based
mobility. It is possible to support mobility for IPv6 nodes without host involvement by
extending MIPv6 (RFC3775) signaling messages between a network node and an HA. This
approach, which supports mobility, does not need the MN to be involved in the exchange
of signaling messages between itself and the HA. A proxy mobility agent in the network
Optimization and Enhancement of MIPv6 in ATN 31

performs the signaling with the HA and does the mobility management for the MN
attached to the network. This protocol is known as Proxy Mobile IPv6 (PMIPv6) just for
the use and extension of MIPv6 signaling and HA functionality.
Network deployments for support mobility would be agnostic to the capability in the
IPv6 stack of the nodes that it serves. Enabling PMIPv6 protocol functionality in the net-
work supports. IP mobility for nodes that have mobile IP client functionality in the IPv6
stack and those nodes that do not. The merits of developing an MIPv6-based mobility
protocol are as follows:

1. Reuse of HA functionality and the messages/format used in mobility signaling.

The protocol of MIPv6 is mature and many implementations have been subjected
to interoperability testing.
2. All types of IPv6 nodes would have a common HA which serve as its mobility
agent.

Readers interested in a network-based mobility protocol solution may refer to RFC4830 for
further documentation.
The network-based mobility management protocol called Proxy Mobile IPv6 (PMIPv6)
is a suggested solution being actively standardized by the IETF NetExt Working Group
and is starting to attract much attention among the telecommunication community and
Internet community. Unlike the various existing protocols for IP mobility management
such as MIPv6, which are host-based approaches, a network-based approach such as
PMIPv6 has salient features and is expected to expedite the real deployment of IP mobility
management.
The fundamental foundation of PMIPv6 is based on MIPv6 in the sense that it extends
MIPv6 signaling and reuses many concepts such as the HA functionality. However,
PMIPv6 is designed to provide network-based mobility management support to an MN in
a topologically localized domain. Therefore, an MN is exempted from participation in any
mobility-related signaling, and the proxy mobility agent in the serving network performs
mobility-related signaling on behalf of the MN. Once an MN enters its PMIPv6 domain
and performs access authentication, the serving network ensures that the MN is always
on its home network and can obtain its HoA on any access network. That is, the serving
network assigns a unique home network prefix to each MN, and conceptually this prefix
always follows the MN wherever it moves within an PMIPv6 domain. From the perspec-
tive of the MN, the entire PMIPv6 domain appears as its home network. Accordingly, it is
needless (or impossible) to configure the CoA at the MN.
Despite the fact that network-based mobility management is considered a better solu-
tion than host-based solution, the progress of completing the detailed scheme is quite late.
Moreover, there are some problems about flow mobility support.
Enhanced flow mobility support is based on the virtual interface in the MN. Virtual
interface hides all physical interfaces from the network layer and above. A flow interface
manager is placed at the virtual interface, and a flow binding manager in the LMA is paired
with it. They manage the flow bindings and are used to select proper access technology
to send the packet. Flow mobility procedure is divided into three cases, which are caused
by new connection from the MN, the LMA’s decision, and the MN’s decision, respectively.
An HNP update request/acknowledge message is defined for LMA decision-based flow
mobility. According to the concept of network-based flow mobility, MN-derived flow han-
dover needs the approval of the LMA.
32 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

2.6  NEMO and NEMO Route Optimization Analysis

Providing transparent and unperturbed Internet service is important to mobile hosts and
has been studied in the IETF for many years until now. A common scenario we can easily
imagine is the emergence of mobile networks, namely, a set of hosts that move conjointly
as a unit, such as on flying aircrafts or any other moving vehicles. The protocols for mobil-
ity support therefore need to be extended from supporting an individual mobile device
to supporting an entire mobile network. Bauer has carefully compared several technolo-
gies that support IP mobility and pointed out that NEMO is the most suitable protocol for
ATN/IPS mobility.9
In this section, we discuss the state-of-the-art NEMO support and also give some
research achievements published in recent years, especially focused on ATN environ-
ments. First, we give some basic knowledge of NEMO and then motivate the problem
by considering typical aeronautical/space scenarios and identify the characteristics that
require new solutions. We then study the design requirements of the protocols that sup-
port NEMO. After that, we review several current approaches for NEMO support and
discuss their advantages and weaknesses (i.e., RO) in addressing the design requirements.

2.6.1  Introduction of NEMO and Its Relationship with MIPv6

The growing use of IP devices in portable applications has created the demand for mobil-
ity support for entire networks of IP devices. NEMO solves this problem by extending
mobile IP. Devices on a mobile network are unaware of their network’s mobility; however,
they are provided with uninterrupted Internet access even when the network changes its
attachment point to the Internet. IETF has already created a set of NEMO protocols to pro-
vide basic NEMO functionality on both IPv4 and IPv6 and continues to work with other
projects to develop advanced performance and functionality-enhancing features. A first
set of NEMO implementations are already available on several platforms including BSD
variants, Linux, and Cisco Systems routing equipment. (The simulator NS-3 has already
given a NEMO simulation solution; see Section 2.3.3 for details.)
Figure 2.7 depicts an example of a NEMO scenario. The router within the NEMO that
connects to the Internet is called the mobile router (MR). (In ATN/IPS, the MR is usually
mounted in the aircraft.) NEMO should reside in its home network when it is not travel-
ing in any other foreign networks. Several addresses blocks are allocated by the home
network, and then the address of the mobile network is assigned to the address block
which is known as the mobile network prefix (MNP). When NEMO is away from home
network, it should keep the assigned address. These addresses will lose the means of
network topological structure when they are not in home network. When some data
packages are being sent to a node in NEMO (this node is named by mobile network node,
MNN), and the NEMO is in foreign network, at this time, the packages will be first sent
to home network. MNN can be any host that always moves with other MNNs or a MR
and forms a whole. The MR will exchange a prefix table with HA and will work like
any mobile IP host. Based on the exchanged prefix, the HA will forward received pack-
ages to MNNs that have the same common prefix. We should note that the MR needs
to be assigned an address by its visited foreign network, i.e., CoA which has been dis-
cussed before. This architecture can be used to transmit and receive packages between
the MNN and CN.
Optimization and Enhancement of MIPv6 in ATN 33

1 Handover

Old NEMO 3 New NEMO

tunnel tunnel
Access Access
network A network B
Home agent
Home network

2 NEMO signaling

NEMO
signaling 4 New path for
payload traffic

CN

FIGURE 2.7
NEMO handover signaling between mobile router and home agent.

As described by Bernardos et al.10, the CN can be treated as a host that is situated in

any position of a network. Assume it will communicate with an MNN by some IP data
packages, the signal flow can be described as follows.

• A IP data package will be transmitted from CN to MNN(A). The IPv6 address of

MNN(A) is affiliated with the MNP of NEMO and is contained in the header of
this IP data package.
• As mentioned above, the IP data package will be first be routed to the home net-
work of NEMO rather than be directly transmitted to MNN(A). In the home net-
work, a particular node called HA will repackage this original IP data package to
another new IP data package. Then the new package will be sent to the MR accord-
ing to its CoA; in this stage, the source address is the HA. An important point is,
from the perspective of CN, that NEMO is always in its home network, i.e., the
mobility of NEMO is transparent to MNN(A) and CN, and already established
connections will not be corrupted while MNN(A) is moving. Most routers have
the function of ingress/egress filtering, so repackaging is necessary to prevent
the situation in which the source IPv6 address is incorrect or is incompatible with
network topology and abandoned by the router.
• After the MR receives the repackaged IP data, it will peel the outer header, then
forward the inner IP data package to MNN(A)
• In the reverse transfer stage, i.e., from MNN(A) to CN, the IP data package gener-
ated by MNN(A) is first packaged by MR and is transmitted to the HA of NEMO.
Secondly, the HA will redirect the IP data package to CN.

In order to distinguish how different types of nodes move in the network topology, MNNs
can be divided into three different categories, as described in the following
34 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

• Local fixed node (LFN): This type of node does not change its point of attachment
in the mobile network while maintaining ongoing sessions. The LFN can be a host
or a router and obtains its IPv6 address from an MNP of NEMO.
• Local mobile node (LMN): This type of node is located in the home network, but
the address of the home network belongs to the mobile network. The LMN can
change its point of attachment while maintaining ongoing sessions.
• Visiting mobile node (VMN): This type of node has its own home network and
implementation of MIPv6 protocol, and it can be an MN or MR. Thus, maintain-
ing ongoing sessions will not affect the VMN’s attached point. A VMN can “visit”
a foreign mobile network, which means the VMN can obtain an address from an
MNP.

In a practical environment, a common scenario known as nest NEMO occurs when a

mobile network is located in another mobile network. It often has the topology that an
MR attaches to an MNN in another mobile network, as described by Bernardos et al. An
example in civil aviation is an airline passenger connected to the in-plane Wi-Fi by his
laptop through MR2 in the ceiling, while MR2 is also treated as an MNN because it is con-
nected to the wireless aeronautical data link, which transmits all flight data (e.g., engine
sensors) to the ground station.
Readers are encouraged to Review the work of Bernardos et al.10 for more details and
illustrations.
On the ingress interface (to the mobile network), the MR will advertise one or more pre-
fixes to the MNNs. Although on the surface NEMO appears to be a straightforward exten-
sion to mobile IP, there are several considerations that are still being investigated in IETF
working groups. These issues include how to do NEMO route optimization and several
considerations with prefix delegation and management as it can be seen from Figure 2.7.
One possible implementation of NEMO mimics PMIPv6 in that it is the network access
points that implement the mobility support. When an MN attaches to the access link, the
ground access point will set up a virtual MR that registers the relevant network prefixes
with the HA. All traffic received by this virtual MR for the registered MNPs will be for-
warded across the air–ground link to the connected MN. The MN need not support any
MIPv6 or NEMO signaling and can also auto-configure the IPv6 address if the virtual
MR sends router advertisements for the MNPs across the air–ground link. When the MN
connects to another ground access point, the virtual router is moved by the ground access
network to the new access point with updates to the HA as if the router was physically
present on the MN.
This implementation model allows for MNs with IPv6 stacks that are agnostic of the fact
that they reside on a mobile link and also reduce the protocol overhead for each message
that is sent across the air–ground link.
The mobility signaling for NEMO is similar to MIPv6. As soon as the MR moves to a
foreign network and acquires a CoA, a home registration is performed with the HA that is
located in the home network. The signaling exchange is performed by means of a BU/BA
exchange, protected by an IPsec security association.
Traffic originating from the MNNs is tunneled by the MR to the HA that forwards pack-
ets to their destination, the CNs.
The home network aggregates the MNPs of all its MRs and advertises this aggregate
prefix to other ASs using an interdomain routing protocol. Hence packets originating from
a CN addressed to an MNN that is attached to an MR are routed to the HA located within
Optimization and Enhancement of MIPv6 in ATN 35

MNN MR HA CN

BU
BA

Signaling User data Tunnel

FIGURE 2.8
The signaling of NEMO.

the home network. These packets are then tunneled by the HA to the current CoA of the
MR. The MR decapsulates the packet and delivers it to the destination, the MNN.
The mobility is transparent to both the MNNs and the CNs as only MR and HA are
performing mobility-related tasks. The original packets are also not modified, due to
the IP-in-IP encapsulation that only adds an additional header on the path between MR
and HA.
Finally in Figure 2.8, we give a brief illustration of NEMO mobility signaling and for-
warding of user traffic over bidirectional tunnel. In Figure 2.8, BU and BA signaling are
exchanged between the MR and the HA. User traffic is exchanged between the MNN and
the CN.

2.6.2  Motivation and Scenarios of Using NEMO in Aeronautical Environment

ATN specification lists NEMO in accordance with RFC 3963 (NEMO Basic Support
Protocol) as an option in ATN/IPS for implementation. The motivation is MIPv6 supports
movement of an individual network node. However, there are scenarios in which it would
be necessary to support movement of an entire network in ATN/IPS.
As we know, several NEMO RO mechanisms already exist. In order to encourage the
development of desired functions in aeronautical NEMO RO solutions, consider that aero-
nautical communication is often used in a terrestrial air-ground network or satellite-based
ocean or polar region communication, so some network component such as access net-
work, HA, CN should be used and analyzed by their topology to determine which type of
RO mechanism is most suitable.

• For airline passenger communications (APC) where is would be wasteful of

­bandwidth to have mobility signaling for every individual passenger.
• When there is a common airborne router supporting multiple traffic types
­provided proper security issues can be addressed.

Here are some reasons and analysis to explain why NEMO is selected as the mobility sup-
port protocol in ATN/IPS environment.
Session continuity: The HA provides the airborne router with a constant IP address—
called the home address (HoA)—from its own network. Traffic is therefore always routed
36 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

via the HA that forwards it to the current location of the MN. Mobile network support:
NEMO extends MIPv6 by introducing an MR that has one or several mobile network.
Prefixes (MNPs). These prefixes belong topologically to the home network. End-systems
that attach to the MR configure their addresses based on the MNP advertised by the MR
and can therefore remain mobility agnostic.
Multihoming: The possibility to register several CoAs with the HA is specified.11 In addi-
tion,12 specifies a policy exchange protocol that can be used to set up forwarding rules
for certain traffic flows, taking into account the additional CoA. Detailed traffic selectors
can be used to identify a flow, based on IP or higher-layer protocol fields such as source/
destination address, port numbers, etc. The MR sends its current policy to the HA, which
sets up forwarding to the MR accordingly. This allows for simultaneously routing traffic
flows over different interface, on the routing path from the MR to the HA as well as on the
path from the HA to the MR.
Security: MR and HA perform a mutual authentication between each other based on
IKEv2. This ensures that the HA forwards packets only to the valid MR.
End-to-end delay: NEMO causes suboptimal routing where traffic always traverses the
HA. If the distance between MR and HA increases, the overall end-to-end latency also
increases.
Scalability: The MR signals its current location to the HA that updates its routing
state accordingly. BGP routing tables remain unchanged as the home network is always
advertising an aggregated prefix via BGP that includes all the MNPs. Scalability is there-
fore linear with the number of aggregates, with regard to the number of announced
prefixes.
Applicability to AAC/APC: The protocol stack on end-systems remains unaffected as
traffic is transparently tunneled between MR and HA.
Convergence time is equal to the time it takes the MR to signal the new location to the
HA, who will then immediately forward traffic to the MN’s new CoA.
Efficiency 1: It takes the MR 1 RTT to signal the new location/CoA to the HA.
Efficiency 2: The tunnel between the MR and the HA inflicts an overhead of a full IP
header upon every payload packet.
Support for ground-initiated communications: As it was the case for the IPsec-based
approach, payload traffic is always routed via the HA. As the MR signals its CoA(s) to the
HA, this traffic can be forwarded to the MR’s current location.

2.6.3  NEMO Route Optimization and Its Application on ATN

A disadvantage of the current NEMO protocol is that all traffic between the MNNs
and the CNs suffers from suboptimal routing, since all packets are forwarded through
the HA that is located in the home network. This is in contrast to MIPv6, which does
provide an RO component for establishing a direct routing path between mobile host
and CN.
A significant number of related works13,14 that have proposed RO solutions for NEMO
are available. An evaluation based on a set of requirements is performed in order to
select the most suitable protocol for the safety-related aeronautical communications
environment. RFC552215 describes the requirements and desired properties of RO tech-
niques for use in global-networked communications systems for aeronautics and space
exploration. (RFC5522 is being developed by aeronautical communications experts who
are outside of IETF, such as ICAO, some independent research institutes, or aeronautical
Optimization and Enhancement of MIPv6 in ATN 37

communication standard groups. The latest version of RFC5522 is October 2009.) The
requirements are as follows:

1. End-to-end latency: the number of intermediate nodes on the routing path

between the end-systems in the mobile network and the ground network should
be as small as possible.
2. Single point of failure: a new mobility-specific node for the purpose of RO should
not constitute a single point of failure.
3. Separability: it should be possible to apply RO only to traffic flows that really
require it. For example, RO should only be applied to all traffic originating from
or destined to one specific MNN, while traffic from other MNNs is still routed via
the HA.
4. Multihoming: the RO mechanism must be fully usable if several interfaces are
present. More precisely, it should be possible to forward a route-optimized traffic
flow via a particular interface/access network.
5. Efficient signaling: both the size and the number of individual RO signaling mes-
sages that are exchanged over the wireless path should be kept small.
6. Ground network impact: the amount of support necessary from the ground net-
work in order to provide RO between the MNNs and the CNs should be limited.
7. Security: the mobility entity on the ground with whom RO signaling is performed
must be able to validate the aircraft as proper owner.
8. Adaptability: the RO scheme should not break applications using new transport
protocols or IPsec.

This set of requirements used in evaluation is partially overlapping with the set that was
used for deriving NEMO as the most suitable mobility protocol. The requirement on appli-
cability to Aeronautical Administrative Communication (AAC) and APC has been omitted
as this investigation is focused on supporting safety-related services. The requirements on
routing scalability, convergence time, and support for ground-initiated communications
have been omitted as they are fulfilled by NEMO, irrespective of the selected RO protocol.
Additional NEMO-specific requirements have been added though, such as separability or
adaptability.
In Bauer’s research,14 various different proposals for performing NEMO RO will be
assessed in the following based on the requirements as mentioned before. The RO proce-
dure can be initiated and performed by different network nodes. Consecutively, four types
of RO strategies are defined by the entities participating in RO signaling. These strategies
are shown in Figure 2.9.

• MNN to CN
• MR to CN
• MR to correspondent router
• MR to HA

In Table 3.1 of Bauer’s PhD thesis,14 a summary of the assessment of the four NEMO RO
solution classes with regard to the previously specified requirements is provided.
38 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

RO signaling RO signaling

MNN MR CN MNN MR

Data traffic Data traffic CN

(a) (b)

RO signaling
RO signaling

MR CR
Data traffic MR HA
MNN
CN MNN Data traffic CN
(c) (d)

FIGURE 2.9
Four different NEMO route optimization classes. (a) MNN to CN, (b) MR to CN, (c) MR to CR, and (d) MR to
HA. (Adapted from Bauer, C., Secure and Efficient IP Mobility Support for Aeronautical Communications. Karlsruhe,
Germany: KIT Scientific Publishing, 2013.)

To better help readers to understand the analysis process better, here we give a brief
review of evaluation of the four optimization types of NEMO. For more details, please
refer to Bauer14.

1.
MNN to CN: The MR only relays the router advertisement of its own access router(s)
to the MNNs. Then the MNNs make use of their own Mobile IPv6 functionality
to form a CoA from this advertisement. Then they perform the Mobile IPv6 RO
signaling with the CNs themselves. So in this method, The MR just plays the role
of a supporter.
a. Advantage: Optimized path is the shortest possible one with direct routing
between MNN and CN. It does not have the problem of single point fail-
ure. MNNs can easily decide for which data flows RO should be performed.
Multihoming is also supported
b. Disadvantage: Overhead is heavy, as RO signaling is performed between every
MNN and CN for which an optimized routing path should be provided. The
MN will also have to manage the CoA of the MNN.
2.
MR to CN: MN performs the MIPv6 RO signaling on behalf of the MNN with a
CN. The MR therefore acts as a RO proxy for the MN.
a. Advantage: The end-to-end path is still optimal because packets are being
routed directly on the path between MR and CN. No single point of failure is
introduced. The MR can also make use of multihoming by registering several
CoA with a particular CN.
b. Disadvantage: Security weaknesses. The MR also has to extend the original
packets with additional headers when performing RO on behalf of the MNN,
Optimization and Enhancement of MIPv6 in ATN 39

so it increases the overhead. The integrity of end-to-end flow is corrupted, so

some security protocols may take the RO signal as a attack.
3.
MR to CR: A correspondent router (CR) is newly introduced in this method. It
is located close to the CNs. The MR exchanges mobility signaling with the CR
and constructs a bidirectional tunnel that is used to exchange traffic between the
mobile network and the network served by the CR. The CR therefore acts as a RO
proxy for the CNs.
a. Advantage: If there is no signal global CR, the problem of single point failure
does not exist. Also the signaling overhead has been reduced compared to the
previous protocols (1) and (2). This RO process is transparent to end-to-end
transport and security protocols
b. Disadvantage: Some security risks are inherited, since the signaling procedure
between MR and CR is based on MIPv6 RO signaling.
4.
MR to HA: The concept of having only a single HA has been extended to having
multiple, distributed HAs within the home network. The proposal is called Global
HA-to-HA. The MR binds to the closest HA to achieve RO. The amount of end-to-
end latency reduction depends on the location of the HA.
a. Advantage: The signaling exchanges between MR and primary HA are similar
to the basic NEMO protocol. Global HA-to-HA does not require any mobil-
ity functionality in the CNs or within the CN networks. It not only provides
a high level of security between MR and HA by IP-in-IP tunnel, but also the
end-to-end transport or security protocols remain unaffected.
b. Disadvantage: The single point of failure problem has only been partially
addressed. The amount of end-to-end latency reduction depends on the loca-
tion of the home agents.

As a conclusion, the above methods (1) and (2) are discarded because the complexity is
increased at the end-systems, and also because of large signaling overhead as well as
security risks. Both (3) and (4), illustrated in Figure 2.10, have their pros and cons. The CR
protocol in (3) is the most adequate solution for safety related aeronautical communica-
tions. For non-safety related communications, especially aeronautical passenger commu-
nications, protocol (4) is the only feasible solution.

2.7 Conclusion
As we know the essential characteristics of aircraft are high speed and high mobility, ATN
should conform to these characteristics. How to keep accurate and timely data transmis-
sion is a critical aspect of the ATN system. MIPv6 is the solution for these problems in the
IP layer.
In this chapter, we give an overview of MIPv6 and its extension, which can be used in
the next-generation ATN system. The discussion is started from the infrastructure of gen-
eral mobile communication, then the motivation of MIPv6, MIPv6 protocol, and related
triangular routing problem is presented. Due to some defects of the original MIPv6 proto-
col, four types of extension are introduced, that is, HMIPv6, FMIPv6, PMIPv6, and NEMO.
40 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

HA in Europe
HA in Asia

HA in the
United States

Communication path
MR-HA tunnel
HA-HA signaling
CR
Correspondent
network
(a)

Home network Communication path via HA

Communication path via CR
HA MR-HA tunnel
MR-CR tunnel

CR
CN Correspondent
network
(b)

FIGURE 2.10
(See color insert.) Typical air traffic services communication scenario making use of different route optimiza-
tion protocols. (a) Global HA-to-HA and (b) communication router. (Adapted from Bauer, C., Secure and Efficient
IP Mobility Support for Aeronautical Communications. Karlsruhe, Germany: KIT Scientific Publishing, 2013.)

Since NEMO is very suitable for a whole mobile network moving, we focus on optimiza-
tion of NEMO and its applications.
This chapter is based on the recent published research articles and technical reports; the
readers can catch the essence of mobility of ATN and can conduct deep study about the
knowledge of a certain subsection.

References
1. ICAO. Aeronautical Communications Panel (Working Group I), Manual for the ATN using
IPS standards and protocols, Doc 9896, First Edition, April 2014 (More details about Working
Groups of ACP Panel are available at http://www.icao.int/safety/acp/Pages/).
2. Sampigethaya, K. Future E-enabled aircraft communications and security: The next 20 years
and beyond, Proceedings of the IEEE, 99(11), 2040–2055, 2011.
3. EUROCONTROL. Challenges of growth 2013, Task 7 Report: European air traffic in 2050.
Available at https://www.eurocontrol.int/articles/challenges-growth. June, 2013.
Optimization and Enhancement of MIPv6 in ATN 41

4. Fistas, N., and B. Phillips. Action plan 17—Future communications study—Final conclu-
sions and recommendations report. Digital Avionics Systems Conference (DASC) Proceedings,
2007, p.12.
5. EUROCONROL. Communications operating concept and requirements for the future radio
system (COCR V1.0), Eurocontrol/FAA Report, 177pp.
6. Sajatovic, M., Haindl, B., Epple, U., and Gräupl, T. Updated LDACS1 system specification. SESAR
P15, 2, 2011. More infomation can be found at http://www.ldacs.com/publications-and-links/.
7. Microsoft Corporation. Understanding mobile IPv6. Microsoft Windows Server 2003 white
paper, 2004. http://www.cu.ipv6tf.org/pdf/MobileIPv6.pdf pp 5–7.
8. Choi, H.-Y., Min, S.-G., and Han, Y.-H, PMIPv6-based flow mobility simulation in NS-3, Fifth
International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS),
pp. 475–480, June 30, 2011–July 2, 2011.
9. Bauer, C. and Zitterbart, M. A survey of protocols to support IP mobility in aeronautical com-
munications, IEEE Communications Surveys and Tutorials, 13(4), 642–657, Fourth Quarter 2011.
10. Bernardos, C.J., Soto, I., and Calderón, M. IPv6 network mobility, Internet Protocol Journal 10, 2,
2007. http://www.cisco.com/web/about/ac123/ac147/archived_issues/ipj_10-2/102_ipv6.
html.
11. Wakikawa, R., Devarapalli, V., Tsirtsis, G., Ernst, T., and Nagami, K. Multiple care-of addresses
registration, RFC 5648, October 2009.
12. Tsirtsis, G. et al. Flow bindings in mobile IPv6 and network mobility (NEMO) basic support. IETF
request for comments 6089 (2011). Details can be found at https://tools.ietf.org/html/rfc6089.
13. Bauer, C. A secure correspondent router protocol for NEMO route optimization, Computer
Networks, 57(5), 1078–1100, April 2013.
14. Bauer, C. Secure and Efficient IP Mobility Support for Aeronautical Communications. Karlsruhe,
Germany: KIT Scientific Publishing, 2013.
15. Eddy, W., Davis, T., and Ivancic, W. Network mobility route optimization requirements for
operational use in aeronautics and space exploration mobile networks (RFC 5522). October
2009. More details can be found at http://tools.ietf.org/html/rfc5522.
3
Modern Air Traffic Control Systems

Yun-Fei Jia

CONTENTS
3.1 Introduction...........................................................................................................................44
3.2 Overview of ATC..................................................................................................................44
3.2.1 Evolution of ATC.......................................................................................................44
3.2.2 ATC Overview........................................................................................................... 45
3.2.2.1 Airport Control and Approach Control................................................. 45
3.2.2.2 Local Control.............................................................................................. 46
3.2.3 Components of ATC................................................................................................. 46
3.3 Radar Data Processing......................................................................................................... 47
3.3.1 Format of Radar Data............................................................................................... 47
3.3.1.1 Radar Data Transfer Protocol................................................................... 47
3.3.2 ASTERIX Protocol..................................................................................................... 47
3.3.3 Track Generation....................................................................................................... 48
3.3.3.1 Protocol Interpretation.............................................................................. 48
3.3.3.2 Coordinate Transformation...................................................................... 48
3.3.3.3 Target Correlation...................................................................................... 49
3.4 Conflict Detection and Warning......................................................................................... 49
3.4.1 CDW Overview......................................................................................................... 49
3.4.2 Short Term Conflict Warning.................................................................................. 50
3.4.3 MSAW......................................................................................................................... 50
3.4.4 Dangerous Area Infringement Warning............................................................... 50
3.4.5 Special Code Warning.............................................................................................. 51
3.5 Flight Plan Processing.......................................................................................................... 51
3.5.1 State Management.................................................................................................... 51
3.5.2 Flight Profile.............................................................................................................. 52
3.5.3 Handover................................................................................................................... 53
3.6 Integral Surveillance and Control...................................................................................... 53
3.6.1 Overview of ISC........................................................................................................54
3.6.2 Tracks and Maps....................................................................................................... 55
3.6.3 Browse and Modify Flight Plan.............................................................................. 56
3.6.4 Record and Replay.................................................................................................... 57
3.6.5 Handover................................................................................................................... 60
3.6.6 Low-Latitude Airspace Infrared Supervision....................................................... 60
3.7 Conclusions/Outlook........................................................................................................... 62
References........................................................................................................................................63

43
44 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

3.1 Introduction
Airplanes must be safely guided in the sky and on the ground. The objective of the air
traffic control (ATC) system is to make sure airplanes be safely separated as they fly in
the sky, land at or take off from the airport, or move around the airport. In addition, ATC
can enhance the airspace usage and expedite the flow of traffic. This chapter introduces
the ATC system.
The ATC system is the man–machine interface that provides various services to in-flight
airplanes. In many countries, the ATC system is set up at air traffic control centers (ACC),
where controllers issue instructions and advisories to pilots.
ATC is a distributed computer system, including several subsystems. These subsystems
process the radar data, calculate the minimum separation between airplanes, manage the
flight plan, and display the tracks. Usually, each subsystem is running at individual com-
puters, and all the computers are connected via a switch, forming a distributed computing
system. These computer systems are connected by a dual network. This design can prevent
network failure. In addition, each computer runs a real-time operating system. This guar-
antees the ATC’s response to the request of a pilot within a certain period of time.
This chapter focuses on the operational principle of the ATC system. We first provide an
overview and discuss the evolution of ATC, including its composition and how it works.
In Section 3.3, we describe the radar data processing (RDP) subsystem, including interpre-
tation of the radar data package, ASTERIX protocol. This section focuses on the process
of track generation. The conflict detection and warning (CDW) subsystem is described in
Section 3.4, including short-term conflict and warning, minimum safety altitude warning,
dangerous area infringement warning, and special code warning. Section 3.4 elaborates
the principle of the design of CDW. Flight planning is the core part of ATC, as is illustrated
in Section 3.5. State management of the flight plan is very important for ATC. The flight
profile, which is calculated from flight plan, can estimate the quality of flight. Flight plan
handover is also mentioned in Section 3.5. The Integral Surveillance and Control (ISC)
terminal is a man–machine interactive system. It shows the calculated results of all the
subsystems, including tracks, conflict warning, and flight plan. All these will be described
in Section 3.6. In addition, record and replay is an important part of the ATC system, the
operating mechanism of which is also described. Section 3.7 concludes this chapter.
To sum up, this chapter will make you aware of how ATC works to guide an airplane
safely to its destination airport. The operational principles of the ATC system are also
included.

3.2  Overview of ATC

3.2.1  Evolution of ATC
With the fast development of commercial aviation, air traffic congestion becomes more
and more serious. In order to guarantee the safety of more and more flying airplanes and
improve the efficiency of airspace usage, the ATC system is designed to guide airplanes
according to a set of predefined rules. Air traffic control refers to the activity of controlling
and/or guiding the airplanes so that they can fly to their expected destinations safely.
Modern Air Traffic Control Systems 45

ATC has undergone four stages of development, which can be roughly described as
follows: Before the 1930s, airplanes flew only in daytime, and the pilots obeyed a set of
rules for “visual flight.” Not long afterward, controllers were employed to guide the
landing and taking off of airplanes with red and green flags. Also, a control tower
was set up at the highest point of the airport, where the controller could monitor the
airplanes on the ground through visual means. From 1934 to 1945, visual flight rules
hardly met the requirement of increasing flight activities. During this period, radio
communication and navigation devices were mounted on airplanes and around the air-
port. In addition, some air traffic administrations such as towers, local ACCs, and flight
stations were set up across the country. More specifically, towers can guide the move-
ment on the airport ground, monitor departure and arrival, and local control centers
can guide airplanes as they are cruising on the air route. The local control centers are
responsible for collecting the position of airplanes, and the controller can determine
the relative position of airplanes and thus direct the flight by communicating with the
pilots. From 1945 to 1980s, radar technology, which was developed during World War
II, was widely used for ATC. The controller can see on the screen the position, call
number, height, speed, and related parameters of in-flight airplanes. Another important
technical advancement is the instrument landing system (ILS), which enables airplanes
to land in bad weather. The fourth stage began in the 1980s, where satellite communica-
tion technology and the Internet are being extensively applied to ATC. With the help of
satellites, the ATC system is united with the airplanes in the air, thus processing com-
munication automatically.

3.2.2  ATC Overview

The primary task of ATC lies in controlling and managing the whole process from take
off to landing. In this process, ATC must guarantee that the airplane will not conflict with
another airplane, or with barriers on the ground. Also, ATC must expedite air traffic and
enhance airspace usage [1,2]. Classified by the controlled airspace, ATC can be divided into
airport control, approach control, and local control.

3.2.2.1  Airport Control and Approach Control

Airport control is usually responsible for guiding the landing process, movement on the
ground, and departure. The controllers in the tower manage and control the airplanes on
the ground mainly by visual means. A tower is a tall building with many windows located
on the airport grounds. Tower controllers are responsible for the separation and efficient
movement of airplane and vehicles on the runways. In addition, some busier airports
may use surface movement radar to help the controller to efficiently manage airplanes.
Consequently, there are usually several displays in the control tower, which can help the
controller to guide the airplanes on the ground. The display may include a map of the local
area, the position of various airplanes, meteorology information, and so on.
Approach control intends to guide incoming cruising airplanes to gradually lose alti-
tude for landing, and outgoing airplanes to gain altitude until they reach the cruising
altitude. Usually, airport control and approach control are carried out by different con-
trollers, yet in the same tower. Only in a busier airport is approach control part of indi-
vidual ACC.
Approach control employs radar to monitor the airplane and guide it via radio commu-
nication. It uses the same surveillance technology as local ATC.
46 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

3.2.2.2  Local Control

After an airplane takes off and climbs to the cruising altitude, local control centers will
take care of it. Because airplanes often fly across a wide area, there are several local control
centers set along the air route. Each will serve the airplane in their respective airspace.
According to the flight plan of an airplane, the controllers in the local control center will
either accept or reject the airplane handed over from the previous local control center. Then
they guide the airplane via radio communication and radar, so that there is a safe distance
between airplanes. Finally, the controller will hand over the airplane to the next local con-
trol center. This process is repeated until the airplane reaches its destination airport.

3.2.3  Components of ATC

The ATC system is a distributed computer system. Each subsystem runs on an individual
computer. These computers are connected in a LAN via a switch. All these subsystems are
connected via TCP/IP protocol. Figure 3.1 illustrates the structure of an ATC system.
The flight plan subsystem will receive and then store in database the flight plan from
airlines. ISC can obtain these plans by querying the database.
The surveillance data from primary radar, secondary surveillance radar (SSR), or ADS-B
are sent to this LAN. The first receiver is the radar data processing subsystem, which inter-
prets the radar data and extracts trace points from them. Then, it translates the collected
trace points to tracks. This work is completed by determining whether the trace points
represent the new positions of the existing track or a new aircraft.
RDP outputs the tracks that include, but are not limited to, an ID and positional infor-
mation about an airplane. Then the tracks are sent to the CDW subsystem, where they are
compared against neighboring airplanes or barriers on the ground to decide whether an
alarm should be triggered. Then the CDW subsystem sends these tracks to the ISC subsys-
tem, which is a GUI application. The tracks will be shown on the ISC.
In addition, the reliability and availability of an ATC system is very important. Hence,
double-network structure is usually employed to avoid network failure, as is shown in
Figure 3.1. Certainly, a highly reliable and available system should adopt several fault-
tolerance techniques. However, this chapter is intended to illustrate the operational
principle of ISC instead of reliability design of ATC.
All the subsystems are elaborated in the following sections.

Radar data Conflict detection

Flight plan
processing and alarm

A A
B B

Integral surveillance Integral surveillance

and control(I) and control(II)

FIGURE 3.1
(See color insert.) Composition of ATC.
Modern Air Traffic Control Systems 47

3.3  Radar Data Processing

Radar data processing (RDP) will receive data collected by several radars and transmit
them to a uniform target track usually by means of data fusion. This process includes
interpretation of a radar data package (may be from various radar manufacturers).

3.3.1  Format of Radar Data

3.3.1.1  Radar Data Transfer Protocol
Transmission protocol for radar data belongs to the data link protocol in the OSI reference
model, including asynchronous protocol and synchronous protocol. Synchronous proto-
col includes BISYNC and HDLC, and the latter is the de facto protocol used in the aviation
industry. HDLC protocol is described as follows.
High-Level Data Link Control (HDLC) is a bit-oriented data link protocol, used to trans-
fer radar data on a synchronous network. HDLC protocol is developed by extending IBM’s
Synchronous Data Link Control (SDLC) protocol. The frame structure of HDLC is illus-
trated in Figure 3.2.
In Figure 3.2, the F1 field denotes the starting point of the frame. Moreover, it is used
to fill certain characters between two frames. Field A represents the address of the source
radar. Each source station has a unique address. However, an address can represent a set
of stations, which is called station set. This field can be set to 11111111 to represent the
broadcast address. This field can also be set to 00000000 to be used for testing purposes.
In Figure 3.2, C represents control field, which is a key field of the HDLC frame. This field
is used to construct various commands and responses for the purpose of surveillance and
control of data link. For example, the sender can use this field to tell the receiver to execute
specified commands. Also, the receiver can use this field to reply the execution status of
the instructions to the sender. The first and second bits in this field represent the type of
transfer frame, including information frame, surveillance frame, and unnumbered frame.
The fifth bit of this field is P/F bit, representing polling. The information field is denoted
by I in Figure 3.2. This field contains the data to be sent, the size of which depends on the
size of the buffer of the computer. Usually, the size is from 1000 to 2000 bits. The last field
FCS represents the frame check sequence, which usually adopts a 16-bit CRC algorithm.

3.3.2  ASTERIX Protocol

Depending on the manufacturer of the primary radar, the format of radar data can be dif-
ferent, including ASTERIX, CD2, MP2, TOSHIBA, MH/4008-2000, etc. The most commonly
used format is ASTERIX. An ASTERIX radar package may contain one or more data blocks.
A data block is composed of category of data (CAT), length of data block (LEN) and a certain
number of records. The data block can be described by Table 3.1.

8 bits 8 bits 8 bits 8 bits*n 16 bits 8 bits

F1 A C I FCS F2

FIGURE 3.2
Frame structure of HDLC.
48 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 3.1
Format of ASTERIX Data
CAT LEN Data Record (1) Data Record (2) ……
1 byte 1 byte Variable-length byte Variable-length byte ……

3.3.3  Track Generation

The primary purpose of RDP is to translate the collected trace points to tracks. A track is a
time series, representing a series of history positions of an aircraft. Each track has a unique
track ID. Trace point is the flying object observed by radar. After each rotation, the radar
collects many trace points. Each trace point may represent a new aircraft or new position
of an existing track. This process is called track generation.
In a scanning period, when receiving new radar data, RDP will usually do the following
tasks as shown in Figure 3.3.

3.3.3.1  Protocol Interpretation

Because the received data may be collected by radars from different manufacturers, they
usually use different transmission protocols as mentioned in Section 3.3.1. Hence, RDP
first interprets the radar data package. More specifically, RDP  extracts the positional
information, speed, and the secondary surveillance radar (SSR) code from the radar data
package.

3.3.3.2  Coordinate Transformation

Radar records the position of airplanes in polar coordinates for convenience, while the
maps and air lines usually employ rectangular coordinates. Hence, we should transform
the positional information, included in radar data packages, to rectangular coordinates.
Moreover, the positional information included in radar data will be corrected. This is
because the location, the sea level at which the radar is located, will result in error of
observation. Hence, each radar data will be corrected depending on the reasons for obser-
vation error.

Radar data Target No Establish new

correlation track

Yes
Data Update original
interpretation Track list
track

Coordinate
transformation

FIGURE 3.3
Track generation.
Modern Air Traffic Control Systems 49

3.3.3.3  Target Correlation

In an observation cycle, radar will collect several trace points. These trace points may
correspond to existing track, or a new one. If it belongs to an existing track, then it will
be marked with an existing track ID. Accordingly, in the next display update cycle, ISC
updates the position of an existing track with this ID. Otherwise, it will show a new track.
Target correlation aims to determine whether a new trace point corresponds to an exist-
ing airplane, or a new one. This is done by comparing the position, speed, altitude of
incoming trace point against nearby existing tracks.
When target correlation is completed, the generated track will be added to the track list
and sent to the CDW subsystem.

3.4  Conflict Detection and Warning

3.4.1  CDW Overview
The first task of ATC is to ensure safe flight of airplanes. The ATC system should warn
the controller when there is any potential danger or harm to flight activity. Accordingly,
the controller can take measures to eliminate or reduce the harm or potential risk. CDW is
designed to automatically evaluate the potential riskful conditions or harm based on the
surveillance information.
The CDW subsystem includes several types of warning, that is, short-term conflict
warning, minimum safety altitude warning, dangerous area infringement warning, spe-
cial code warning, and so on.
Conflict warning is judged by the CDW and shown on the ISC. If there is any conflict
warning, various warning signals will be shown on ISC. ISC shows flickering track plates
in different colors. Usually, a yellow flickering track plate denotes a less serious risk situa-
tion and red flickering track plate indicates dangerous conflicts. In addition, ISC warns the
controller with a warning sound. Conflict warning is illustrated in Figure 3.4.
Once the controller gets a warning signal, they can roughly judge the cause of conflict
and quickly identify the track in question. Then, according to the causes of conflict, the
controller can guide the aircraft to eliminate the conflict condition. If the conflict condition
is eliminated, the warning signals will gradually disappear. Typical conflict conditions are
described in the following sections.

36 CLAM 36 STCA
CSN6201 M CSN6201 M
0056 V 0090 031 0056 V 0090
CAUC A319 CAUC A319

FIGURE 3.4
(See color insert.) Conflict warning.
50 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

3.4.2  Short Term Conflict Warning

Short-term conflict warning (STCW) will warn the controller of possible conflicts between
airplanes. The ATC system will issue an STCW when the horizontal interval and the verti-
cal space of flight-pair are less than the minimum safety distance.
STCW will calculate the interval of each flight-pair. Thus, when there are N airplanes
flying in the airspace of concern, the number of flight-pairs (denoted by M) can be calcu-
lated by

N ⋅(N − 1)
M = CN2 = .
2

From this equation, we can obviously see that when there are too many airplanes in the
airspace, the number of flight-pairs can be astronomical. This will overload the ATC sys-
tem, leading to inability to output real-time warnings. For simplicity, we can use filtering
methods to reduce the computing time. Typical filtering methods include mosaic filtering,
coarse filtering, and fine filtering. In this book, we only introduce the mosaic filtering
method.
The mosaic filtering method divides the horizontal project of the airspace into a number
of mosaics, and each mosaic includes a number of tracks. Because the flying speed is a
limited value, two mosaics far from each other will not result in short-term conflict. Thus,
we can filter some flight-pairs for further processing.
It should be noted that the mosaic filtering method only considers the horizontal space
between two tracks. Hence, the length of the side of a mosaic should be well considered.
If it is too long, there will be too many tracks in a mosaic, and the filtering effect will be
distorted. If the size is too small, there will be too many mosaics in the airspace of concern,
resulting in poor searching performance.

3.4.3 MSAW
A minimum safety altitude warning (MSAW) is designed to detect the interval between
airplane in the air and the obstacles on the ground. Obviously, this type of warning detec-
tion depends on the topographic information of the airspace of concern. For simplicity, we
can adopt a square mosaic, the side length of which equals the highest point of the topog-
raphy of concern, to represent the minimum safety height.
Under some conditions, MSAW should be restrained. For example, when an aircraft is
taking off or landing, there is very small space between the aircraft and the buildings
on ground. In this case, the false warning signal will disturb the work of the controller.
MSAW can be restrained by the controller by specifying the area where MSAW is not
calculated.

3.4.4  Dangerous Area Infringement Warning

A dangerous area infringement warning (DAIW) is issued when an airplane is close to
or infringing some special area. These special areas include restricted airspace, danger-
ous airspace, and so on. Dangerous areas can be temporal or long-term. Some long-term
dangerous areas are specified beforehand by the controller. Also, there are some temporal
dangerous areas. For example, when there are military manoeuvres, that airspace will be
restricted for a period of time. Temporal dangerous areas are also specified by the controller.
Modern Air Traffic Control Systems 51

Dangerous areas are defined by a polygon column, the top surface of which is parallel
to the ground. A metric denoted by the minimum safety distance is defined in advance. If
the distance between track and that predefined polygon column is less than the minimum
safety distance, a DAIW will be issued to the controller.

3.4.5  Special Code Warning

Besides these three types of warnings, some special warnings are also issued. In extreme
cases, the pilot will press the emergency buttons on the instrument panel in the cock-
pit. Accordingly, the ATC will receive some special codes from the pilot, including 7500
(hijacking), 7600 (onboard radio devices failure), and 7700 (airplane in emergency). Under
those conditions, the ISC will warn the controller.

3.5  Flight Plan Processing

The flight plan is a detailed description of the coming flight activity. A piece of flight
plan includes airplane identifier (usually airline number and task number), flight rules
(instrument flight, visual flight), flight category (general aviation flight, military flight,
non-periodical flight, periodical flight), airplane number, type of airplane, type of wake
(heavy, medium, or light), onboard devices, departure airport, expected departure time,
cruising altitude, destination airport, expected departure time, alternate airport, second-
ary alternate airport, etc.
The flight plan is established by the aviation operation units (usually airlines), and sent
to air traffic administration. When the ATC system receives a piece of flight plan, it will
store it in its database and manage its state. Also, the received flight plan can be modified
at the ISC by the controller.
Flight plans can be long term and temporal. In commercial aviation, most flight plans
are relatively constant. Long-term flight plans are stored in the ATC system. Thus, the air-
line does not need to send a flight plan to ATC, unless it is changed.

3.5.1  State Management

Flight plan processing (FPP) will take charge of the state transition of a flight plan, includ-
ing automatic acceptance of flight plan, storage of flight plan, management of state,
assignment of SSR code, correlation of flight plan with track, forecasting track, and track
handover.
State of flight plan includes inactive state, preactive state, coordinated state, controlled
state, handed-over state, and finished state [3]. Figure 3.5 shows the state transition of a
flight plan in chronological order.
When FPP receives a piece of flight plan, it sets it to the inactive state. Before a period
of time the airplane takes off (usually half an hour), the flight plan will be set to the pre-
active state. Just before the airplane takes off, the air traffic controller will communicate
with the pilot to determine if everything is OK. Then, the airplane will take off and the
flight plan will transit to the coordinated state. After the airplane takes off, the radar
will capture it and correlate it with its flight plan, then the flight plan will transit to the
52 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Inactive Controlled

30 min before taking off Set by controller

Preactive Handed-over

Confirmed by
2 min after landing
pilot and controller
Correlated with
track
Co-ordinated Finished

FIGURE 3.5
State transition of flight plan.

controlled state. Because an airplane usually flies a great distance, several air traffic cen-
ters are located along the flight line. When the airplane flies past an air traffic center, the
controller will set the flight plan to the handed-over state. In such a situation, both the
current air traffic center and the next air traffic center will see the airplanes on the screen
of the ISC. After the controller at the next air traffic center accepts the flight plan, it will
transit to back to the controlled state. Usually, after a period of time (usually 2 min) from
the airplane’s landing at destination airport, the flight plan will automatically transit to
the finished state. At this point, the lifetime of the flight plan is over.
Civilian airplanes are usually identified by an SSR code, which is a 4-bits octal num-
ber. Thus, there will be 212 = 4096 SSR codes at most. In fact, some codes are used only
for special purposes, so the available SSR codes for civilian airplanes are very limited. In
countrywide airspace, there may be more airplanes than the limited number of SSR codes.
The solution is to assign a unique SSR code to a civilian airplane in a certain control area.
While the airplane flies to another control area, the assigned code will be changed. The
rationale behind this solution lies in that, in a control area, there are a limited number of
flying airplanes and each can be assigned with a unique code.
Another important function of FPP is to correlate track with the corresponding flight
plan. This is implemented through the SSR code assigned to the airplane by the controller
in a certain control area. Further, the control center in that control area will use a secondary
surveillance radar to supervise the airplane, the ATC in that control center will receive the
secondary surveillance radar data that contains an identical SSR code to that mentioned
earlier. Thus, the track can be correlated with the flight plan.

3.5.2  Flight Profile

FPP can be used to forecast when the airplane arrives at each way point and by what
route. Further, the flight profile can be estimated, which includes the time and altitude of
the airplane arriving at each way point. This information is extracted from the flight plan.
A possible flight profile is illustrated in Figure 3.6.
A complete flight process can be divided into three stages: climbing stage, cruise stage,
and descent stage. The airplane will depart the airport and then try to climb to cruise
Modern Air Traffic Control Systems 53

Approaching
Cruising and landing
Take off
Climbing

FIGURE 3.6
Flight profile.

altitude. The airplane will fly at the cruise altitude usually at constant speed. Finally,
the airplane will descend steadily when it approaches the destination airport. It should be
noted that, in Figure 3.6, the height of the starting point and the end point of the profile
is usually different. This is because the altitude above sea level of the departure airport
and the destination airport will be different. In the flight plan, we can obtain the code of
departure airport and thus its height above sea level. Based on the airplane model field in
the flight plan, we can estimate the climbing rate and descent rate. Also, with the expected
departure time field and cruise altitude in the flight plan, we can draw the curve at the
climbing stage. The curve at the cruising stage can be easily drawn, because the expected
time to arrive at each way point (denoted by E1, E2, E3, and E4) can be directly obtained
from the flight plan, and remember the cruise speed is constant. At the descent stage the
curve can be drawn in the same way as that at the climbing stage.
Flight profile can also be used to calculate the expected flight-related parameters (flight
route, speed, etc.) based on the approved flight plan, and warn or prompt the controller if
the actual flight parameters are inconsistent with those in the flight plan.
In addition, a flight profile can show the planned track on the ISC. This is meaningful
when the aircraft is flying in the blind zone of the primary radar. Under this condition, the
track will disappear on the ISC and the planned track (denoted by a different tag) will be
shown. When the aircraft has flown across the blind zone, the track will be shown again
and the planned track will disappear.

3.5.3 Handover
In two cases, the FPP will hand over the flight plan. In the first case, when airplane flies
across control areas, the FPP will transfer the control to the next control area. In the second
case, when the airplane flies across control sectors, the FPP will transfer it to another con-
troller. In both cases, FPP should change the state and send necessary flight plan informa-
tion to the receiver. Handover is elaborated in Section 3.6.5.

3.6  Integral Surveillance and Control

ISC is a man–machine interactive system, which is an important part of the ATC system.
The main functions of the ISC are threefold: (1) to display the supervision data from the
primary radar, infrared camera, or ADS-B; (2) to process the flight plan and present it in
the form of a flight strip; and (3) to exchange information with the controller and the pilots.
54 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

In an ATC, there may be more than one ISCs cooperating with each other to complete the
task. In Figure 3.1, two ISCs are employed, that is, ISC I and ISC II.

3.6.1  Overview of ISC

The main window of the ISC is depicted in Figure 3.7, which can be roughly divided into
three fields: the formation field on the top toolbar, man–machine interaction field in the
middle part, and the function field at the bottom toolbar.
The features in the format field can be illustrated as follows. We can set the System mode
to “NORM MODE” or “REPLAY MODE.” As its name implies, NORM MODE means the
ISC is currently used to monitor the air situation, while REPLAY MODE refers to the ISC
being used as a terminal to replay the air situation. REPLAY MODE is usually used for
accident analysis or newbie training. The measurement units are either imperial units or SI.
Users can select the time zone in the UTC time field. An alarm clock can be used to remind
the controller about important matters at specific time intervals. The longitude-latitude
field shows the cursor position.
The man–machine interaction field lies in the middle part of the main application win-
dow. This is the main field to show and/or manipulate the tracks, flight trip, and various
maps such as section maps, landmark maps, border maps, and so on. For example, users
can select any flight trip and edit it, or select any track to change the style, or measure the
distance between two tracks. The concentric rings shown in Figure 3.7 help the controller
to determine the distance between two points.
The function field at the bottom toolbar provides most of the functions of the ISC.
Clicking any button in this field opens a pop-up window with more details. Clicking the
“Map” button shows each map layer. The “Flight Plan” button can be used to set or edit the
flight plan in detail. The “Drawing” button is used to select drawing tools to draw maps
on screen. This field also provides a record and replay function, an infrared surveillance
function, etc.

FIGURE 3.7
(See color insert.) Overview of ISC.
Modern Air Traffic Control Systems 55

In this section, the five modules of the ISC will be described, that is, track and maps,
flight plan, infrared surveillance, record and replay, and track handover.

3.6.2  Tracks and Maps

On the screen of ISC, a real-life airplane is denoted by a hexagon symbol. This symbol
is referred to as a track in this section. It is linked with a piece of string, called a track
plate. The track plate shows detailed information about the flying airplane, such as the
altitude, position, international registration number, and so on. The ISC uses different
symbols to represent tracks to clearly differentiate the data source. The data from radar
is indicated by the symbol “+” and from ADS-B by the symbol “△.” When manipulating
the track, users should select the track first. When the track is selected, a red hexagon is
set around it, and the corresponding flight strip is also selected. A typical track and its
track plate are shown in Figure 3.8.
From Figure 3.8, we can see some dashed dots following the track that represent the his-
torical positions of this airplane in a period of past time. The controller can optionally set
how many history trace points will be shown. History trace points can help the controller
to determine the airplane’s point of origin.
Usually, the ISC includes three types of maps, system map, online map, and generated
map. System maps refer to those relatively static maps, such as airspace map, section map,
air line map, guidance station map, airport map, and so on. These maps are designed as
layers. Thus, user can display any number of maps arbitrarily. If necessary, these maps
drawn by the controller can be released to other ISC terminals, when two or more ISC
terminals are employed, as shown in Figure 3.9. For example, temporally risky area maps
are often released to other ISCs.

FIGURE 3.8
(See color insert.) Track.
56 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

FIGURE 3.9
(See color insert.) Maps.

3.6.3  Browse and Modify Flight Plan

If these airplanes have submitted a flight plan, they will get useful information from the
ATC system when they are flying, such as meteorology, military activity area, ground
proximity alarm, and so on. [4]. Particularly, the airplanes flying a commercial route will
get beneficial information. A flight plan window is shown in Figure 3.10.
The ISC flight plan window provides rich information about the flight process, mainly
including type of airplane, expected time of departure, expected time of arrival, destination

FIGURE 3.10
(See color insert.) Flight plan window.
Modern Air Traffic Control Systems 57

FIGURE 3.11
(See color insert.) Flight strip.

airport, etc. Moreover, if the in-flight airplane modifies its flight plan, this flight plan win-
dow can update the information in real time.
There are primarily four types of flight-plan-related manipulations: flight plan browse,
flight plan modification, flight plan state change, and flight plan correlation. There are two
ways to browse a portion of the flight plan: (1) select a track, then open flight plan win-
dow, which displays the flight plan of a particular track, and (2) directly input the flight
number into the flight plan window to obtain the details. In some cases, an airplane might
change its intention of flight when it is flying. For example, when an airplane changes its
destination airport, flight route, or expected time of arrival due to bad weather, its flight
plan should be modified accordingly. The controller can edit fields as shown in Figure 3.10
to modify the flight plan and synchronize it with the database. Flight plan state change
means it changes to the preactive state from the inactive state, or to the active state from
the pre-active state, and so on [3].
Flight plan correlation means the correspondence between the flight plan and track.
It should be noted that a track plate will not be shown around the track unless the track
is correlated with the flight plan. In most cases, the controller will frequently observe the
flight plan. Thus, a compact version of a flight plan called the flight strip is created, which
is always shown on the screen. A flight strip is illustrated in Figure 3.11.
It should be noted that all the information shown in a flight strip can be found in the flight
plan window. In Figure 3.11, all fields of the flight-plan-related information are arranged in
a specific order, and each field represents some information extracted from the flight plan.

3.6.4  Record and Replay

Record and replay is a powerful technique to analyze the cause of an air accident. Record
and replay will record the surveillance data, flight plan, operations of the controller, etc.
When air accidents occur, we can replay the air situation to analyze the cause and inves-
tigate the responsibility of the controllers. Also, the feature of record and replay is often
used for training newbies [5].
The design objectives of record and replay include (1) low overhead during recording, (2)
interactive operation during replay, (3) low disk usage, and (4) sufficiently robust technol-
ogy. Low overhead indicates that the recording process cannot affect the operation of the
ISC or consume too many computing resources. Interactive operations are required when
replaying the recorded air situation and operations of controllers. It is meaningful for acci-
dent analysis. For example, users can measure the distance between two airplanes when
58 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

replaying the recorded operations. The recorded files cannot occupy much disk space, as
long-term recording will exhaust disk space. The recording and replaying process must be
implemented by a reliable and robust technology. Some methods to implement a record
and replay system are as follows:

Screen capture: This method captures the screen in a video file. All operations of the
controller and air situation are recorded. This method is simple to implement.
However, it has several drawbacks: (1) The recording requires too much CPU
usage. (2) The recorded video files are too large, so a common PC cannot store
recorded files for very long. (3) When replaying the air situation, the controller
cannot intervene. For example, the controller may want to measure the distance
between two airplanes or the distance from airplane to ground when replaying
an air accident.
Windows hook: This method takes advantage of the message-driven mechanism of
Windows OS. All the messages processed by the ISC can be hooked and recorded
in a file. When replaying the air situation, we can open the file and get the mes-
sage and its time, then send the message to the ISC. The advantage of this method
is obvious. The recorded file is small and the overhead of the recording process is
relatively small. However, this is a complex and error-prone method. In principle,
the ISC responds to the stored Windows message and reexecutes the command
of the controller and shows the recorded track data from radar or ADS-B data. If
there is any error in the stored messages or track data, critical failure in the ISC will
be inevitable. The disadvantage lies in that this is not a robust way to redisplay
an air situation. In addition, we cannot replay the air situation from arbitrary time
instances by using this method. Moreover, this is a platform-specific method.

Another method is using the Xwindow protocol of Unix OS. The recording application
uses Xwindow protocol to store the state of application in Xserver. This is a platform-specific
method like the Windows hook method, and we also cannot replay the air situation from
arbitrary times.
Obviously, these schemes cannot meet our objectives. Thus, we do not adopt these meth-
ods to implement the record and replay module of the ISC system in this chapter. Instead,
we record the original operational data of the ISC, including the following four parts: (1)
operation time; (2) operation type, including window-related operations, track-related
operations, and track plate-related operations; (3) target UI object, including the IDs of
controls and its parent controls; and (4) user’s input; this refers to the results applied to the
ISC. For example, the input in an edit box control will be recorded. A sample of recorded
contents is summarized in Table 3.2.

TABLE 3.2
Recording Sample
Operation of Controller Recorded Content
Main window operation IDs of controller
Child window operation IDs of controller and its parent window
Select/deselect track Position of track in the linked list, flag bit
Modification of track plate Position of track in the linked list, modified value
Modification of flight plan International no., IDs of modified controls
Online drawing Color, width and line shape and positions of each vertex
Modern Air Traffic Control Systems 59

We record the operations of the controller in the manner of incremental recording, that
is, if the controller does not operate for a period of time, nothing is recorded. This method
can significantly save disk space.
We conduct four experiments to test the efficiency of our approach against others in our
ATC platform. We specially test the disk usage of each solution, because CPU usage test
is only required when there is a large overload, that is, when too many operations need
to be recorded. This occurs rarely in general aviation. In our test, a typical air traffic man-
agement process is run on our ATC. Two airplanes fly from one place to another in four
rounds.
The four rounds cost about 20, 40, 80, and 160 min, respectively. During the flying, the
operations of controller are recorded by each method, and disk usage is recorded respec-
tively. The test results are shown in Figures 3.12 and 3.13.
From Figures 3.12 and 3.13, we can see that disk usage in our method is greatly less
than others. In addition, the curve in Figure 3.13 shows a flattening trend. This is to say
our method utilizes less disk space when the flight period increases. This is due to our
incremental recording design as mentioned earlier. This can be interpreted as follows: Any
flight will require specific interactions with the controller, such as departure report, flight
plan manipulation, landing report, and so on. Usually, the controller will rarely intervene
in the flight plan, except in dangerous situations. As the traveling time increases, the idle
time of controller will increase. The idle time refers to the period when the controller per-
forms no operations.
Several fault-tolerance techniques are implemented in the record and replay module to
improve its robustness. The duration of the recorded files does not exceed 5 min. In this
way, not too much information is lost even if faults occur during replaying. Moreover, we
record the “global state” of the ISC at the head of each recorded file, including the state of
all windows, controls and maps, tracks, etc. The recorded global state can be treated as a

500
* Windows hook
450 + Screen capture

400

350
Disk usage (kB)

300

250

200

150

100

50

0
20 40 60 80 100 120 140 160
Time (min)

FIGURE 3.12
Disk usage of other solutions.
60 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.5

6.5

6
Disk usage (kB)

5.5

4.5

3.5

3
20 40 60 80 100 120 140 160
Time (min)

FIGURE 3.13
Disk usage of our solution.

savepoint. If any faults occur during the replaying process, we can roll back to the save-
point then retry. These measures can significantly enhance reliability.
Considering the interactive operation during the replay of an air situation, this method
allows the controller to intervene in the replaying process. For example, the controller can
measure the distance between two airplanes, zoom in/out on the screen, etc. These opera-
tions can help the controller to make an in-depth investigation of the cause of air accidents.

3.6.5 Handover
When an airplane flies very long distances, it will be controlled by several air traffic centers
until it reaches the destination airport. Hence, the airspace is divided into many controlled
zones, each controlled by an air traffic center. When the airplane flies to the boundary
of two air control zones, the ATC system will automatically detect this and the track on
the ISC will start to flicker. It should be noted that, the primary radars set in both control
zones will see the airplane, and the controllers in both air traffic centers will see the flick-
ering tracks. Usually, the next controller will manually accept the airplane by pressing the
“ACC” button on the menu pop-up by right-clicking the flickering track. In fact, the han-
dover process is to transit the jurisdiction of the flight plan from the current ATC center
to the next one.

3.6.6  Low-Latitude Airspace Infrared Supervision

In low-latitude airspace, general aviation also should be guided. For general aviation, ADS-B
is the primary communication technique between controller and pilot. The advantage of
Modern Air Traffic Control Systems 61

ADS-B is two-fold: (1) it is very economical; and (2) it can provide rich and accurate informa-
tion about the airplane, such as the speed and operating state of the airplane. Nevertheless,
the disadvantage of ADS-B is obvious. Because ADS-B uses a one-way broadcasting tech-
nology, the receiver need not respond to the requests from the ADS-B. Hence a hostile
flying object carrying ADS-B can disguise itself as a common airplane by sending fake
datagrams. For example, if an invasive fighter sends false datagrams to the receiver of the
ATC, the controller cannot identify it. It should be noted that the primary radar can only
detect but not identify the flying object. More specifically, based on the available infor-
mation, the primary radar can only get the position and speed of the flying object, but
not what it is.
Thus, we need other surveillance techniques to identify unidentified flying objects.
Infrared surveillance can show the shape, size, flying altitude, alpha, position, etc. [6]. In
sensitive areas such as border area, infrared surveillance can be used as a complementary
surveillance technique for the ATC.
Infrared surveillance systems include an infrared camera that is usually mounted with
the primary radar. Because an infrared camera cannot endure long-term operation, it is
in hot standby state. When the primary radar detects an unidentified object, the infrared
camera will start recording automatically [7]. With the positional information from the
primary radar, the infrared camera can easily capture the flying object.
The rationale behind the integration of the primary radar and infrared camera lies in the
fact that: first, we use primary radar to detect flying objects and infrared camera to iden-
tify them; second, infrared camera cannot locate the flying object at startup state without
the assistance of the primary radar; and finally, the infrared camera cannot endure long
working times—most of the time it is ready but not working. The integration of both will
make the detection and identification better.
When the infrared camera finds the flying object, it will capture it on video. The video
stream will be transferred to the ISC by network. The dataflow in a typical infrared sur-
veillance is shown in Figure 3.14.
Figure 3.14 illustrates the collection, transmission, and display of an infrared video stream.
There are several coding standards for video coding and compression. More specifically,
H.263 is usually used for video meeting and MPEG-1 and MPEG-2 for multimedia applica-
tions. In this chapter, we use H.263 coding standard due to its real-time property. Finally,
the infrared video stream will be received and shown by ISC, with which the controller
observes and identifies the flying object by visual means.
The infrared surveillance window of ISC is shown in Figure 3.15, in which a flying air-
plane is captured by an infrared camera. In Figure 3.15, we can clearly see the shape of the
flying object.
It should be pointed out that, Figure 3.15 only shows a screenshot of the video. In fact,
the infrared surveillance window can display the full video. As mentioned earlier, when

Video Coding and IDC

Network
stream compression

Infrared camera

FIGURE 3.14
Infrared video transmission.
62 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

FIGURE 3.15
Infrared display window.

starting, the infrared camera needs the help of the primary radar to locate the suspected
flying object. Once the object is located, the infrared camera can automatically track
the flying object by means of some image recognition.

3.7 Conclusions/Outlook
In this chapter, we introduce the modern ATC system from the perspectives of the user
and developer. The composition of the ATC system is illustrated, and some peripheral
equipments such as ADS-B, secondary surveillance radar, and surface movement radar
are also mentioned. Some design experiences are presented in this chapter, including the
considerations when selecting a better method.
All subsystems of the ATC system are described in the order of data flow. First, we
describe how to extract positional information from a radar data package and translate
it to track. Then the positional information will be used to calculate various conflicts.
Management of the flight plan is illustrated. Finally, surveillance data display and ground-
to-air communication are described in detail. In addition, we introduce the design and
implementation of the record and replay module, including how to increase the perfor-
mance, availability, and reliability.
To sum up, the operating mechanism of the ATC system is clearly understood. After
reading this chapter, you will know how to design an ATC system and how to use it to
control in-flight airplanes.
Modern Air Traffic Control Systems 63

References
1. Li, R.X., Bar-Shalom, Y. Design of an interacting multiple model algorithm for air traffic control
tracking. IEEE Transaction on Control System Technology, 1, 186–194, 1993.
2. Dimitris, B., Lulli, G., Odoni, A. An integer optimization approach to large-scale air traffic flow
management. Operation Research, 59, 211–227, 2011.
3. Zhang, W., Kamgarpour, M., Sun, D., Tomlin, C.J. A hierarchical flight planning framework for
air traffic management. Proceedings of the IEEE, 100, 179–194, 2012.
4. Everdij, M.H.C., Scholte, J.J., Blom, A.P., Stroeve, S.H. An investigation of emergent behaviour
viewpoints in literature, and their usability in Air Traffic Management. 2011. http://reports.nlr.
nl:8080/xmlui/bitstream/handle/10921/144/TP-2011-​444.pdf?sequence=1. Accessed March
28, 2014.
5. Curtis, S., Hallett, E., Mirchi, T. Training air traffic controllers for future next generation air
transportation system (NextGen) technologies. California State University, 2014. http://chaat.
cla.csulb.edu/research/Publications/Curtis%20Hallett%20Mirchi.pdf. Accessed March 28,
2014.
6. FAA (Federal Aviation Administration). 91-36D—Visual Flight Rules (VFR) flight near noise-
sensitive areas. 2004. http://www.faa.gov/regulations_policies/advisory_circulars/index.
cfm/go/document.information/documentid/23156. Accessed March 28, 2014.
7. Li, J., Gong, W. Real time pedestrian tracking using thermal infrared. Imagery Sensors, 5(10),
1597–1605, 2010.
4
Security of VHF Data Link in ATM

Meng Yue

CONTENTS
4.1 Introduction........................................................................................................................... 66
4.2 ACARS System Principle and Message Format............................................................... 66
4.2.1 Overview.................................................................................................................... 66
4.2.2 Research on the Composition and Operation of ACARS................................... 67
4.2.2.1 Composition of ACARS............................................................................. 67
4.2.2.2 Operation Mode of ACARS...................................................................... 68
4.2.2.3 How the Airborne Management Unit Works........................................ 69
4.2.2.4 How the Ground Network Works........................................................... 69
4.2.3 ACARS Message Format.......................................................................................... 70
4.2.3.1 ARINC Specification 618 and the ACARS Air–Ground Message
Format.......................................................................................................... 70
4.2.3.2 ARINC Specification 620 and the ACARS Ground–Ground
Message Format.......................................................................................... 73
4.2.4 Conclusion................................................................................................................. 76
4.3 ACARS Datalink Message Security Protection Method................................................. 76
4.3.1 Security Threats of ACARS Datalink.................................................................... 76
4.3.1.1 Data Leakage..............................................................................................77
4.3.1.2 Data Deceiving........................................................................................... 78
4.3.1.3 Entity Masquerade..................................................................................... 78
4.3.1.4 Denial of Service........................................................................................ 78
4.3.2 ACARS Message Security........................................................................................ 79
4.3.3 ACARS Information Security Architecture..........................................................80
4.3.3.1 DSP-Based Security Architecture............................................................80
4.3.3.2 End-to-End Security Architecture........................................................... 81
4.3.4 AMS Implementation Scheme and Key Technology........................................... 82
4.3.4.1 Data Encryption.........................................................................................83
4.3.4.2 Message Authentication Code..................................................................84
4.3.4.3 Digital Signature Scheme......................................................................... 85
4.3.4.4 Key Management....................................................................................... 86
4.3.4.5 Payload Encoding...................................................................................... 88
4.3.4.6 Information Coding................................................................................... 88
4.3.4.7 Data Compression...................................................................................... 90
4.3.5 Conclusion................................................................................................................. 91
References........................................................................................................................................ 91

65
66 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

4.1 Introduction
Digital air–ground communication allows transfer of digital information between equip-
ment onboard an aircraft in flight and ground-based computers connected through a data
link service provider’s (DSP’s) transport network. There are many ways to support the data
link, such as line-of-sight very high frequency (VHF), high frequency (HF), and satellite.
As data link use increases, it becomes necessary to have data link channels constantly
available.1
The Aircraft Communications Addressing and Reporting System an (ACARS) is a data
link system that allows communication of character-oriented data between aircraft sys-
tems and ground systems. It enables the aircraft to operate as part of the airline’s com-
mand, control, and management system.
Equipment required onboard for data link is known as an ACARS management unit
(MU). The MU is connected to a standard airborne transceiver for data link communication
and may be connected to other airborne equipment via an airborne communication bus.
The communications management unit (CMU) provides ACARS communications function
in lieu of an ACARS MU. The CMU is designed to support bit-oriented message transfer
utilizing the protocols and procedures for routing. The CMU will interact with onboard sys-
tems that supply mobile air–ground communications over VHF, satellite, and HF media.2
There are two parts in this chapter, the ACARS principle and message format are
described first and the ACARS data link message security protection method is detailed
in the second part.

4.2  ACARS System Principle and Message Format

ACARS, as an important means of current aeronautical communication system, is widely
used in aviation and is essential to ensure flight security. This section will detail the com-
position of ACARS, how it works, and the ACARS message format.

4.2.1 Overview
ACARS is a character-oriented data link, which cannot transmit digital voice and data
stream files, such as satellite cloud images. This system consists of an airborne subsystem
(management unit and control unit), VHF remote ground station (RGS), ground data com-
munication network, ground network management and information processing system
(NMDPS), and data link users.3
Over the years, people always want to provide the latest information to the aircrew
through the voice communication system without adding an extra burden. ACARS
enables the aircrew to send information to the route command center, such as departure
time, arrival time, fuel situation, departure delay, and other information. The ACARS sys-
tem remains aircrew workload tractable and provides the aircrew and route command
center with information services at the same time. Other advantages of ACARS include
ground surveillance capabilities on aircraft engine and other parameters, a more effective
exchange of information related to flight arrival and sustained flight, reducing multiple
frequency changes in aircraft, and a more reliable selective call system.
Security of VHF Data Link in ATM 67

At present, the main international data link service providers include Aeronautical Radio
Incorporated (ARINC) (United States), Societe Internationale de Telecommunications
Aeronautique (SITA) (Europe), and Automatic Data Direction Control (ADCC) (China),
and they have different data link service area coverage. Each airline can choose a different
DSP to provide services. When the aircraft flies through different service areas, it needs to
deliver information through the international gateway.

4.2.2  Research on the Composition and Operation of ACARS

This section details the composition of the ACARS data link communications system and
describes how ACARS works.

4.2.2.1  Composition of ACARS

ACARS consists of airborne subsystems and ground stations network. The ACARS air-
borne subsystem is composed of VHF transceiver, VHF, VHF antenna, control unit, MU
and ACARS control unit (CU), as shown in Figure 4.1. The ground station network is com-
posed of VHF ground station, a central processing computer and a conversion network
connected with the computers of each route.
ACARS is a collaborative system providing air–ground and ground–air digital voice
and data communications. The ACARS network provides a common data link that can
handle all kinds of messages. The ACARS control management unit enables airborne sys-
tems transmit and receive work.4
ACARS is an air–ground communication network, which can connect aircraft as a
mobile terminal with the route command, control, and management system. Information
about the departure and arrival time, departure delayed information, fuel conditions,
and other flight-related data is automatically collected and sent to the system through the
control unit. The system uses International Organization for Standardization (ISO) letter,
fifth character, and this format is fully compatible with the agreement of the United States
Air Transport Association (ATA) and the International Air Transport Association (IATA).

Communication
satellite
VHF

RGS RGS RGS

Ground
network
Users
Network management and
data processing center

FIGURE 4.1
(See color insert.) The composition of ACARS.
68 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Maximum block length is 220 characters, and a message longer than 220 characters will be
divided into a number of blocks.
The message format consists of seven units in accordance with the International Civil
Aviation Organization (ICAO) conventions of encoded characters. Eighth is coupled with
a character to provide parity bit.
Information is sent from the plane transmitted via 131,550  MHz ACARS data link to
terrestrial radio stations, and then it is relayed to a central computer processor where the
information is converted into a route-task message. The route-task message is sent to the
relevant routes by ARINC electronic switching system (ESS).

4.2.2.2  Operation Mode of ACARS

ACARS has two work modes: request mode and polling mode.5 Request mode allows air-
borne- or ground-processor platform to initiate communication. When there appears a
scheduled event or a ready command input by the pilot, airborne systems will be auto-
matically ready to launch. Then, after determining the ACARS channel without traffic,
the management unit will launch the news. If there is traffic on the ACARS channel, the
system will wait before transmitting until the message traffic is eliminated. If two or more
airborne systems choose the same moment to transmit, the messages will be chaos. In
order to avoid transmission synchronization, the messages in the future will be reemitted
at random intervals.
As soon as the ground station processor receives a transmission message, it completes
the error test of block check sequence (BCS). If the message is errorless, the processor will
send it to its proper destination. Ground station processor also sends a confirmation signal
<ACK>, which can notice the airborne to eliminate the original message and return to the
normal work. If the message is wrong, the processor does not send the confirmation signal
and airborne systems will reemit messages. If the message has not been confirmed after
six times, the system will warn the aircrew. If the uplink transmission (ground-to-air) is
wrong, the downlink message sends a negative signal <NAK>.
The polling mode only allows the ground processor to initiate communication. During
the polling mode, airborne ACARS can only respond to the received uplink message
(polling). Once the airborne system is in the polling mode, the ground processor main-
tains a continuous communication with airborne system by transmitting a general polling.
When the radio channel is unblocked, the airborne system will respond to these polling,
or, if no message, it will reply with the general response of the polling mode. BCS error
detection is performed in all messages (uplink and downlink) and generates a confirma-
tion or a denial message (ACK or NAK) included in the next transmission. Ground proces-
sor sends a general response of request to command the airborne systems switch to the
request mode. When ON event (landing) appears, airborne system work in polling mode
will automatically enter the request mode.
For all sampling data during the course of scheduled flights, MU has played the role of
organizer or format converter. The management unit collects the data from control units,
aircraft sensors, and out, off, on, and in (OOOI) events sensor. The OOOI event sensors
determine the flight arrival, departure, and flight time. Management unit generates a
Greenwich Mean Time clock to record the occurrence time of OOOI events. At the request
of the ground processor, the flight information, cumulated by cockpit flight management
computer (FMC), integrated data systems, and aircraft flight recorder terminal, is format-
ted as a digitally encoded signal and is transmitted to ground processor via VHF trans-
ceiver machine.
Security of VHF Data Link in ATM 69

All uplink (ground–air) emissions are monitored by the management unit to determine
the address. The message’s destination address is identified through the aircraft registra-
tion number or flight number. Aircraft registration number is identified by the unique
lapping form of aircraft wiring harness. Flight number is entered into the memory man-
agement unit by the flight crew before the flight. Unless the address of registration number
or flight number matches the aircraft name, the management unit will not respond to any
uplink message.
ACARS uses 131 and 550 MHz horizon VHF radio frequencies according to the current
provisions. Additional frequencies can be added as the demands increase. ACARS in HF
frequency range is in favor of the extension of air digital services when aircrafts fly over
the ocean.

4.2.2.3  How the Airborne Management Unit Works

The management unit is the processor for receiving data from a number of sources. They
must also be easy for future expansion when data link system functions increase. So, the
management unit is basically a central computer or processor with a variety of other data
processor interfaces. These data processors are subordinate to the main processor and do
not conflict with the unit, so that the main processing function can compile all the flight
data and format them.
The main processor uses the bus arbitration logic to communicate with the slave pro-
cessor. This kind of signal exchange control form makes main processors read data from
the secondary processor and send data to the main processor memory. In addition, the
main processor also generates a continuous Greenwich Mean Time for internal and exter-
nal clock and communicates with the control unit. Monitor and keyboard information
is received via the serial input bus and stored in the data buffer. Once the data buffer is
full, the main processor will do the error checking of the display information. In order to
control the display of the control unit, the main processor compiles the data and transmits
it through the serial bus to the control unit. Display information containing specified posi-
tion for fault management unit or control unit advertised.
The input–output processor monitors OOOI event input and accesses overlapping infor-
mation from the aircraft wiring system. This information includes the aircraft registration
number, route identification, OOOI sensor leads, and program management unit program
leader. The input–output processor is also used to send Greenwich Mean Time information
to the control unit for display or to the serial output bus for external use in data link system.
The modulation–demodulation processor is used as the interface between data link system
and VHF communication system. The modulation–demodulation processor is composed of
a demodulator, and the processor must check whether parity uplink message and address
(valid flight number or aircraft identification) are correct, and whether the BCS code exists.

4.2.2.4  How the Ground Network Works

A downlink transmission from an onboard ACARS unit is demodulated at the radio ter-
minal and stored in the local memory buffer. This downlink transmission asks each radio
terminal buffer at the central processor AFEFS (ACARS front end processing system)
whether they concentrate the received message. Message structure is the same as the mes-
sage structure of the radio link. The message format between AFEFS and ESS adopts the
agreement of  ATA and IATA. Each message consists of the necessary addressing informa-
tion, routing information, and free text (if used). The message structure is arranged by the
70 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

standard message identifier (SMI) and the body of the identifier (TEI). AFEFS completes
all the functions related to link control, message formats, message queuing, and network
monitoring. AFEFS both receives and transmits messages from the ESS.

4.2.3  ACARS Message Format

Data link communication uses the air–ground network to transmit a certain telegram
meeting format in the space. The ground–air communication format should conform to
ARINC Specification 618 and the ground–ground communication format should conform
to ARINC Specification 620.

4.2.3.1  ARINC Specification 618 and the ACARS Air–Ground Message Format
• The main content of ARINC Specification 618
• ARINC Specification 618, also known as air–ground character-oriented protocol
specification, includes ACARS system description, air–ground block structure
and message handling protocols, link management for a VHF air–ground net-
work, the air–ground satellite link protocol, and HF data protocol for transporting
character-oriented ACARS messages. The ACARS message is introduced accord-
ing to the definition of the ISO 5 character set.6
• The air–ground block format.
• The format of the air–ground message is governed by ARINC Specification 618.1
The downlink block format is as shown in Table 4.1. The uplink block format is as
shown in Table 4.2. The message consists of header, text, suffix, and BCS. The header
includes start of header (SOH), mode, address, technical acknowledgment, label,
downlink block identifier/uplink block identifier (DBI/UBI), and STX (start of text).

The text field of all messages must consist of the noncontrol characters of the ISO 5 char-
acter set. The maximum length is 220 characters. Longer messages, known as multiblock
messages, will be divided into several separate ACARS messages to send.
For a single block message, its trailer is with the end of the text (ETX). For multiblock
messages, the trailers of all blocks are with the control character end of block (ETB), except
that the trailer of the final block is with the ETX.
The BCS is the result of a cyclic redundancy check (CRC) computation to ensure that the
block is error free. The BCS is initiated by, but does not include, the SOH character, and is
terminated by, and does include, the ETB or ETX character. The BCS is generated on the
entire message, including parity bits.7

1. SOH: 1 character
This character is represented by the control character SOH in ISO 5 character set.
An SOH character (ASCII code value “0 × 01”) indicates the start of the message
header. SOH is also used to indicate the beginning of the text. This text is evalu-
ated by BCS but should not be checked.
2. Mode character: 1 character
The Mode characters are divided into two basic categories as follows:
Category A—This category is denoted by a “2” character. Aircraft may broadcast
the message to all DSP ground stations in the aircraft’s VHF coverage area.
TABLE 4.1
ARINC618: The General Format of Air–Ground Downlink Messages
Name SOH Mode Address TAK Label DBI STX MSN Flight ID ApplText Suffix BCS BCSSuffix
Security of VHF Data Link in ATM

Size 1 2 7 1 2 1 1 4 6 0~210 1 2 1
Example <SOH> 2 .N123XX 5Z 2 <STX> M01A XX0000 DOWNLINK <ETX> <DEL>

TABLE 4.2
ARINC618: The General Format of Ground–Air Uplink Messages
Name SOH Mode Address TAK Label UBI STX ApplText Suffix BCS BCSSuffix
Size 1 2 7 1 2 1 1 0~220 1 2 1
Example <SOH> 2 .N123XX 10 A <STX> UPLINK <ETX> <DEL>
71
72 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Category B—Aircraft may transmit the message to a single DSP ground station in
the aircraft’s VHF coverage area. The address of ground station is placed in Mode
character position and is represented by <@> or <]>. The actual Mode character
inserted in each downlink message is determined by the ground system access
code with a prefix. This access code is selected by MU/CMU based on the uplink
Mode character.
3. Address: 7 characters
It is the aircraft registration mark (tail number). MU/CMU should not transmit
any downlink message having a valid aircraft registration mark. Once the CMU
acquires the aircraft registration mark, it should record it until power to the CMU
is removed. The address can identify the aircraft communicating with the aircraft
processor. In air–ground messages, the Address field should contain the aircraft
registration mark. The aircraft registration mark may be provided by wiring at the
MU’s interface within the aircraft (typical of characteristic 597 and 724 MU’s) or it
may be provided in the form of ARINC 429 data words by an external subsystem
(typical of the ARINC 724B MU).
4. Acknowledgment: 1 character
The positive technical acknowledgment character consists of an A–Z, or a–z,
or a NAK control character. It is the acknowledgement for the message. The
sender needs an acknowledgement from the recipient after sending a message.
ANRINC618 provides a dedicated acknowledgment reply message to confirm a
message that has been received. The technical acknowledgment message contains
Label characters DEL. If the recipient needs the feedback of some types of pack-
ets from the sender after receiving the message, it can set TAK field to < ACK >,
showing that the receiver message is received successfully. When the sender learns
that the previously sent message was successfully received, it needs to confirm the
message with <ACK> from the recipient, hence reciprocating, until one of them
uses <_DEL>. Generally, the TAK character in the first message of a complete com-
munication is <NAK>, and the following is <_DEL> message or a message con-
taining some information with TAK character <ACK>. If it is <_DEL>, the TAK
field of the received message will be <BI> field.
5. Label: 2 characters
This Label is in the message preamble. The Label character is used as a shorthand
method to describe the message. The Label field indicates the classification of the
content of the message being transmitted, and it is used to determine routing and
addressing. A list of assigned labels is defined in ARINC specification 620, such
as Out/Return IN Report (IATA Airport Code)—Label QG, and IN/Fuel Report
(IATA Airport Code)—Label QD. The corresponding meaning and the format of
different categories of message are not the same.
6. Downlink block identifier/uplink block identifier (DBI/UBI): 1 character
ARINC618 agreement provides that the <BI> field of the current message being
sent cannot be the same as the <BI> field of the previous message. The uplink
block identifier (UBI) should consist of an “A”–“Z” or “a”–“z” or single NUL
character. The downlink block identifier (DBI) character should consist of a 0–9
character.
Security of VHF Data Link in ATM 73

7. STX: 1 character
The control character STX indicates the end of preamble and the beginning of
the text.
8. Text: Not exceeding 220 characters in length
The Text field of all messages must consist of the noncontrol characters of the ISO
5 character set. The maximum length is 220 characters. Longer messages, known
as multiblock messages will be divided into several separate ACARS messages
to send. A message sequence number (MSN) and flight identifier (FI) must be
included in the downlink message. The MSN is a four-character field used for
ground-based message reassembly. The FI consists of a two-character Airline
Identifier and a four-character Flight Number field. MU/CMU should not trans-
mit any downlink message having a valid aircraft identification. There are subla-
bels and addresses included in some downlink messages used to address.
9. Suffix: 1 character
Each single block and the last block of multiblock should be terminated with the
control character ETX. The other blocks of multiblock should be terminated with
the control character ETB.
10. BCS: 16 bits
The BCS is the result of a CRC computation to ensure that the block is error free.
The BCS is used for the characters from SOH character to ETX or ETB character,
but does not include SOH character.
11. Control character DEL: 1 character

DEL is BCS suffix, transmitted following the BCS. The purpose of DEL is to enable the last
bit of the BCS to be decoded.
Conclusions from the segment definition are as follows: logic routing and addressing
functions implemented in the message, messages in the downlink realizing restructuring
on the ground, ensuring the accuracy, timeliness and completeness of the message, and
laying a good foundation for the NMDPS to process and distribute message.

4.2.3.2  ARINC Specification 620 and the ACARS Ground–Ground Message Format
1. The main content of ARINC Specification 620
Data link users, aircraft and ground users, use different protocol with DSP. DSP
provides not only the appropriate routing information, but also the message for-
mat conversion between the two kinds of protocol. ARINC620 describes the air–
ground communication protocol and ground–air communication protocol, the
function input and output of DSP systems from the data link user’s perspective.8
ARINC620 agreement mainly involves DSP functions, such as information trans-
fer, the flight tracking, network management, and DSP access.
2. The ground–ground message format
The message processed by DSP can be divided into two categories according to the
flow of information on the data link: uplink message, from the ground user to the
aircraft, and downlink information, from the aircraft to ground user.
74 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

DSP processing information relates to two aspects. DSP converts the downlink
message sent by the aircraft in accordance with ground–ground message format,
and then sends it to the ground user via ground communication network. DSP
converts the uplink message sent by the ground user in accordance with air–
ground message format, and then sends it to the aircraft via data link. Characters
used in ACARS messages are limited to ISO 5 character set.9
1. Downlink ground–ground message format
The DSP receives the downlink messages from the airborne through a ground
station. The downlink message format conforms to ARINC Specification 618. The
DSP converts the received ACARS message according to ground–ground message
format in the ARINC Specification 618, and then delivers the message to the user
via ground communication network.
When the DSP receives a downlink ACARS message from the ground station, the
Label within the ACARS message indicates whether the DSP generates a ground–
ground message and sends it to the ground user. The general format of ground–
ground downlink messages is shown as Table 4.3.8
Line 1, the Priority/Destination Address line, also known as simply the Address
line. The contents of the line consist of two parts: the characters of the priority
of the ground message and the address list of the information recipient. Each
message priority identifier consists of two characters, followed by a space and
then the destination address list including seven characters. The maximum
number of addresses is 16.
Line 2, the address and sending time of the sender. This line begins with the period
character <.> and is followed by the address and a timestamp in the format
day/hour/minute (ddhhmm). It is possible to enter further signature informa-
tion after the timestamp.
Line 3, the SMI line. It contains a three-character code and it is a part of message
type flags.
Line 4, the Text Element field of the message. It is a series of text elements. Each
text element is composed of three parts: text element identifier (TEI), data, and
a text element terminator (TET), as shown in Table 4.4.8 The first text element
is usually the FI, which is composed of a two-character airline identifier and
a four-character flight number. DSP has the responsibility to translate the
three-letter code (as defined by ICAO) airline identifier to the appropriate two-
character IATA equivalent.

TABLE 4.3
General Format of Ground–Ground Downlink Messages
Line Contents Example
1 Priority/destination address QU ADRDPAL
2 Signature/transmission time DSPXXXX 121212
3 SMI AGM
4 Text elements FI XX0001/AN N123XX
5 Communication service line DT DSPRGS121212M01A
6-n Free text - DOWNLINK
Security of VHF Data Link in ATM 75

TABLE 4.4
Text Element Structure
Field Length Coding
Text element identifier (TEI) 2 characters Alpha/numeric
Text element delimiter 1 character Space
Text element data field Variable (depends on TEI)
Text element terminator 1 character </> If another text element follows or
<cr/lf> if this is the final text element

Line 5, the communication service line, contains the text element identifier DT and
the final block. It consists of four fields:
a. DSP identifier. The identity of the DSP provider can be got from this field.
b. If the ground station receiving ACARS downlink messages uses ARINC618
for data transmission, it means this station have received the first block of
the downlink message.
c. Message reception time stamp.
d. Message sequence number as provided in the downlink ACARS message.
Line 6, free text. It is optional and is not part of a message’s structured text.
2. Uplink ground–ground message format
DSP converts the uplink message sent by ground user to the ACARS uplink block,
and sends it to the specified aircraft via ground station.

Uplink ground–ground messages received by the DSP for transmission to an aircraft are
accepted if the following conditions are fulfilled:

a. The timestamp of the message indicates the effective time.

b. The SMI is valid and is approved by the ACARS user.
c. The SMT contains either an aircraft registration number (AN) text element or
a FI text element and the corresponding valid text. If the DSP does not have
tracking information for the aircraft addressed, the following supplementary
condition applies:
i. If the SMT contains either a GL text element (approximate geographic
location of aircraft) or an AP text element (airport location of aircraft) and
an airport or city that DSP can identify, the DSP will determine the ground
station for transmission to the aircraft based on this information.
ii. The application text is preceded by the special TEI—also referred to as
dash space <-sp>, which is used as a separation.
iii. The Line 1 of the ground–ground message must contain the address of
the DSP.

If the ground uplink message sent by the user cannot meet the above conditions, DSP will
intercept the information and send it coupled with the reason of intercepting. The general
format of the ground–ground uplink message is shown in Table 4.5.8
The structure of uplink messages in Table 4.5 is the same with the structure of uplink
messages in Table 4.3. Table 4.4 describes the Line 4 (text elements) in the Table 4.3.
76 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 4.5
General Format of Ground–Ground Uplink Messages
Line Contents Example
1 Priority/destination address QU CTYDPAL
2 Signature/transmission time .QXSXMXS
3 SMI AGM
4 Text elements AN N123/MA 123A
5-n Free text - UPLINK

DSP can get the address of the aircraft in the uplink message contained in aircraft tail
number (AN) and FI. When text includes transmission path (TP), it indicates that the user
designates a transmission path for the uplink message. When the aircraft is not in flight,
DSP can send the message to the specified aircraft via ground station or airport that can
be gotten from a geographic locator (GL) text element or an airport locator (AP) text ele-
ment. When the uplink message includes message assurance (MA), it means that the user
requires the confirmation after DSP delivers the message to the aircraft.

4.2.4 Conclusion
This section introduces the working principle and the structure of the ACARS data link sys-
tem and makes a detailed analysis about air–ground communication format (ARINC618)
and ground–ground communication format (ARINC620) in the ACARS data link.

4.3  ACARS Datalink Message Security Protection Method

Without suitable protection, ACARS datalink communications are vulnerable to threats
posed by unauthorized entities that may access or modify message content, the result of
which may be to expose sensitive information or endanger the safety and integrity of
aircraft operations. This chapter will make a detailed analysis of the vulnerabilities
of the ACARS datalink. Also according to the characteristics of the ACARS datalink, the
practical ACARS datalink message security protection method will be proposed.

4.3.1  Security Threats of ACARS Datalink

The ACARS datalink belongs to the wireless network and compared with wired networks,
it is more vulnerable. All ACARS messages are completely transferred in the form of plain-
text, and anyone can get almost all the ACARS information very conveniently using a
computer, sound card, or radio frequency (RF) antenna and freely available software. Also
the existing software can simulate aircraft or controller terminals to allow calls to the real
terminals. China does not attach importance to the security of ACARS system. There are
no any safety precautions in the air-to-ground VHF link to protect the safety of the link,
but only with a simple user management in the user terminal. This method can only pro-
tect the system from unintentional illegal intrusion. When faced with a malicious attack,
the system will be helpless. So it cannot effectively ensure the security of the system. In a
nutshell, without suitable protection, there are currently a great number of security threats
to the ACARS network.10
Security of VHF Data Link in ATM 77

4.3.1.1  Data Leakage

Existing ACARS datalink security is so poor that any attacker who has an RF transceiver
can intercept ACARS message and understand its contents. ACARS data information
of civil aviation contains a series of airlines-sensitive information, such as take-off and
landing data, oil data, and unit data. If an attacker intercepted and analyzed the data,
the interests of the airline may be directly affected and cause huge economic losses. At
present, more than 70% passenger aircraft in China are equipped with ACARS, and air-
lines are more and more dependent on it to ensure flight safety and significantly improve
the company’s operating efficiency. At the same time, higher demands for the safe and
stable operation of ACARS are put forward. According to the regulations of the Civil
Aviation Administration of China, all aircraft should be equipped with ACARS airborne
electronic devices and applications of ACARS will continuously expand as military air-
planes are also asked to add ACARS airborne communications systems. As military air-
craft have high confidentiality, the consequences of data leakage would be unimaginable.
Currently, ACARS messages sent by planes can be intercepted by using ACARS receiving
software and a simple wireless receiving device. There are common software tools, such
as WACARS and acarsd, which is shown in Figure 4.2.
Data leakage may not directly affect flight safety, but it has become a great hidden trou-
ble for ACARS datalink.11

FIGURE 4.2
(See color insert.) ACARS receiving software acarsd 1.65.
78 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

4.3.1.2  Data Deceiving

Because of the lack of effective security protection, spurious ACARS information could
be transmitted through the ACARS datalink and can appear to be legitimate. Even effec-
tive ACARS information is also very likely to be tampered or resent during transmission,
leading to data error and directly impact on aircraft safety. Now, ACARS messages contain
CRC check mechanism that only allows code error testing in the process of transmission
and does not have the ability to detect man-made deliberate manipulation. Because it is
impossible to confirm message integrity, it cannot be determined whether the message
received is actually from the legitimate sender. Data deceiving will directly affect flight
safety, as the message recipient cannot identify the authenticity of the information.

4.3.1.3  Entity Masquerade

In ACARS, an entity can easily be disguised as a terminal to damage the air-to-ground
communication and to hinder the normal operation of the system. For example, a com-
puter with simple equipment can simulate a controller to send illegal control messages to
the aircraft and it is very easy to cause a catastrophe such as collision, etc.
Entity masquerade can be of two kinds as follows:

1. Masquerade as an aircraft
At present, communications between aircraft and ground receiving equipment
need the aircraft tail number and the service code of airlines, as the basis of DSP
network operation control center to identify the aircraft. Both pieces of informa-
tion can be obtained publicly, and then they can be used to masquerade as an
airplane through relevant electronic equipment.
2. Masquerade as an airline
Currently, the ways that link network provides services to airlines are: first, to
validate the router interface’s IP address of airlines, and then, the system provides
further data services only for the legitimate IP addresses (network layer). Second,
to validate the username of airlines login (application layer); only through the ver-
ification, DSP network control center can finish identity authentication for users,
and then provide users with the corresponding data services. It could be found
that the two security measures are implemented in different network layers.

In the DSP network operation control center, firewall technology has been widely adopted
to strengthen the protection of user data in a datalink network. The DSP network opera-
tion control center only allows specified users to communicate by using the specified
name, specified communication port, and does not open up the other unspecified rights
to users.
Some means cannot fundamentally solve the problem of entity masquerade at present,
and entity masquerade is a serious threat to flight safety and a major hidden trouble of
ACARS datalink.

4.3.1.4  Denial of Service

There are two methods of denial of service attacks in the ACARS system. In one way as
the ground station can only service to one terminal, the attacker can send a lot of false
information to the ground station. As a result, the ground station will fail to respond
Security of VHF Data Link in ATM 79

normally to aircraft communications, bringing about a denial of service to the ground

station. The other way is that the attacker disguised as a terminal sends  lots of useless
ACARS information to the datalink, overloading the information processing center server
and depleting resources, thus resulting in the server’s denial of service. Denial of service
attacks can cause the disruption of normal communication, becoming a serious threat to
the safety of flight.

4.3.2  ACARS Message Security

Aiming at aforementioned security threats of the ACARS system, ACARS message secu-
rity (AMS)2 needs to provide some security services. AMS protects ACARS datalink mes-
sages from unauthorized disclosure and modification, and it provides communicating
entities with assurance that the source of messages is as claimed.
The three main services are as follows:

1. Data confidentiality
Data encryption is the most commonly used method to prevent data leakage.
The AMS data confidentiality service provides users with assurance that the con-
tent of an AMS message is protected from disclosure to an unauthorized third
party during data transmission. In end-to-end mode, in order to guarantee correct
data delivery, its routing information cannot be encrypted, and only the informa-
tion body can be encrypted. In addition, identity authentication can also prevent
illegal intrusion to the system and avoid data leakage.
2. Data integrity and message authentication
Data integrity and message authentication services are treated as a composite secu-
rity service. The authentication mechanisms (e.g., digital signatures and message
authentication codes) that make a recipient able to corroborate the data source also
provide assurance that data was not modified in transmission. In other words, if
the data is authentic, then it cannot have been modified and if the data have been
modified, then it cannot be authentic.
a. Data integrity
The AMS data integrity service provides communicating entities with assur-
ance that an AMS message is protected against modification after the creation
and transmission by an authorized source. Modification includes insertion,
substitution, or deletion of message content, either accidental (such as errors
caused by a noisy transmission channel) or intentional (such as errors intro-
duced by an unauthorized entity).
b. Message authentication
The AMS message authentication is also known as data origin authentication,
and it can effectively mitigate the spoofing threats to provide a message recipi-
ent with assurance that the source of an AMS message is as-claimed. In other
words, message authentication provides the aircraft crew or an automated
aircraft information system with confidence that the messages they received
originate from the claimed ground source, for example, DSP or airline opera-
tions. Similarly, it will provide a dispatcher or automated ground informa-
tion system with confidence that messages they received originate from the
claimed airborne source such as an aircraft information system.
80 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Data integrity and message authentication security services can ensure that the
service objects are authentic and effective terminals, and prevent junk informa-
tion transmitted in the datalink, so it can also effectively prevent denial of service
attacks.
3. Key establishment
The AMS key establishment service provides communicating entities with the
ability to establish cryptographic keys in a secure and authenticated way. The
keys are necessary to support the data confidentiality, data integrity, and message
authentication services.12

4.3.3  ACARS Information Security Architecture

ACARS information security architecture contains two modes: air–ground security archi-
tecture based on DSP and end-to-end security architecture.

4.3.3.1  DSP-Based Security Architecture

The DSP-based security architecture provides security assurance by DSP, relying on the
ACARS ground host and the PKI/CA to distribute the keys. The system architecture pro-
vides a security envelope between an aircraft avionics system and a DSP ground system
that implements AMS, as shown in Figure 4.3.2 The AMS-protected message is protected
during the transmission in the air–ground subnetwork. When the AMS-protected mes-
sage is processed, the original message is recovered, and will be formatted and routed
by DSP. The formatting and routing functions of DSP only apply to existing, nonpro-
tected ACARS messages. So, the DSP may employ non-AMS security services to protect
message exchanges when they traverse the ground–ground network between the DSP
and the ground systems with which the aircraft entity is exchanging messages.

Key Intra-
distribution organization
network
VHF
ATSP

Intra-
organization
network
DSP
Airline

Ground–
Air–ground ground Intra-
network network organization
network
Third-party
Security envelope service provider

FIGURE 4.3
(See color insert.) DSP-based security architecture.
Security of VHF Data Link in ATM 81

It is suitable to use DSP-based security for protection of ACARS messages:

• The DSP must be able to read and process ACARS messages to ensure proper
operation of the air–ground network.
• The DSP provides the formatting and routing of ACARS messages.
• The destination of ACARS message does not implement AMS.

The demand for DSPs is higher in the DSP-based security architecture. They need to provide
not only datalink service to the end users such as airplane, airlines, air traffic control, and so
on, but also undertake the responsibility of PKI (public key infrastructure) at the same time,
being responsible for key distribution and management, providing encryption, and authen-
tication security services. In this security mode, DSP distributes the initialization key (in the
case of using the shared secret key) or digital certificate (in the case of using public/private
key). In communication, a MAC is generated by using a shared secret key between DSP and
aircraft or making digital signatures using an elliptic curve algorithm to initialize security
connection, to complete the basic information exchange and the establishment of a security
connection, and to generate the session keys KENC and KMAC, respectively for ACARS
message encryption and generation of MAC. Key exchange uses the ECDH algorithm (the
elliptic curve Diffie–Hellman), and the session keys are generated by using key generation
function KDF. When the connection is established, the two communication sides verify the
MAC (generated by using hash message authentication code [HMAC] calculation) first,
and then progress or response message after the legal status is mutually determined. In the
DSP-based security architecture, encryption and authentication devices are installed on the
aircraft and DSP respectively. The security envelope is shown in Figure 4.3.
The advantage of this security mode is that safe operation is transparent for operators
(including airlines and the military); it is DSP that centralizes the routing of ACARS infor-
mation; hence, it will be very difficult for the attacker to get routing information in the
air–ground network to make data analysis.
But its drawbacks are also obvious: first of all, DSP must carry out a heavy upgrade to
provide security services. Second, there is no internal information security protection in
the ground–ground network, so an additional mechanism is needed. Furthermore, data
are not secret for DSP, and the plaintexts saved by DSP also become a huge potential safety
hazard, and internal attacks within DSP cannot be prevented.

4.3.3.2  End-to-End Security Architecture

In the end-to-end security architecture, key distribution and management are carried out
by the operating mechanism to ensure the safety of ACARS. The end-to-end security archi-
tecture, as shown in Figure 4.4,2 provides a security envelope between an aircraft avionics
system AMS and an airline or third-party service provider ground system. Both terminals
implement AMS. AMS-protected message exchanges are protected during the transmission,
so the DSP cannot read or process the encrypted content. DSP simply routes the messages
to the intended ground entity according to the information taken from the protected mes-
sage label, airline code, and the airline-specified destination. In this security architecture, the
ground system is also responsible for handling ACARS supplementary addresses contained
in the protected message and for ACARS message formatting performed by a DSP.
It is suitable to use end-to-end security architecture for the protection of ACARS mes-
sages containing airline-proprietary or sensitive information, such as information linked
to individuals.
82 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Key
distribution Intra-
organization
network
VHF
ATSP

Intra-
organization
network
DSP Airline

Ground–
Air–ground ground Intra-
network network organization
network
Third-party
Security envelope service
provider

FIGURE 4.4
(See color insert.) End-to-end security architecture.

However, it is not suitable for the protection of ACARS messages that a DSP must be able
to read and process to ensure proper operation of the air–ground network.
The main difference between the end-to-end and DSP-based security models is that the
former requires the operating mechanism, namely, the airlines, and other end users take
the responsibility of key management and distribution. Implementation means on the
technical side are the same.
The advantage of this model is to provide end-to-end security, to guarantee the secu-
rity of the whole link from the aircraft to the ground host of the operator, namely, the
air–ground link + ground–ground link. Security protection for each session has real-
ized multiple secure sessions with multiple destinations. Encryption is transparent for
DSP and preservation of the ciphertext information by DSP can effectively avoid DSP
internal attacks. However, the operator needs to make large changes to protect safety in
this way.
The end-to-end model can achieve a higher level of security. However, security protec-
tion is needed for each operator itself, which limits its feasibility.
It can be seen that the task of key management and distribution is not taken by profes-
sional security agencies in both DSP-based security architecture and end-to-end security
architecture and there will be problems on safety and feasibility. As an improvement to
the two modes, the task of key management and key distribution can be separated and
handed over to a third party, forming a security framework based on the third party. In
this way, DSP and terminal users can avoid upgrading for a large security framework.
And special security agencies are responsible for key distribution and management that
ensure the feasibility of the solution.

4.3.4  AMS Implementation Scheme and Key Technology

AMS needs to provide the following services to protect the security of ACARS system: link
encryption service, message authentication service, integrity service, and key management
service. And each service can use a different techniques. Link encryption service can use a
Security of VHF Data Link in ATM 83

data encryption algorithm; information authentication and integrity service can use digital
signatures or message authentication codes, and the key management service can use a key
exchange algorithm and key generation function. However, because the resources of ACARS
airborne equipment and the bandwidth of VHF air–ground data link are limited, the appro-
priate technology should be chosen to achieve the best effect in specific implementation.

4.3.4.1  Data Encryption

In general, encryption technology is divided into symmetric and asymmetric encryption.
Symmetric encryption is also called the private key encryption, using the same key to
encrypt and decrypt. Asymmetric encryption is also called public-key encryption with
two separate but related keys, one of which must be confidential as with symmetric
encryption while the other key can be publicly distributed. A comparison of the two kinds
of encryption technology is shown in Table 4.6.
Both encryption techniques can be used to encrypt ACARS datalink. Compared with
public key encryption technology, due to the advantages of low computational cost and
high encryption speed, symmetric encryption technology is more suitable for ACARS,
whose resources and bandwidth are limited. The secret key of symmetric encryption must
be transmitted in a safe manner, and the period key is short, but it can be solved by the
application of key management where the session key is one-time pads.13
In addition, ACARS information is character-oriented, but encryption is generally bit-
oriented. As a result, the ciphertexts that are nonprintable characters probably cannot be
transmitted through the ACARS datalink. This will be introduced in the following coding
section.
CFB (cipher feedback) mode in AES (advanced encryption standard) is adopted for the
reason that the resource and bandwidth of ACARS system are limited. The length of the
key is 128-bit, marked as AES128-CFB128.14 The main reason considered is that the AES
algorithm possesses high safety and CFB mode allows encrypted data with arbitrary
length, convenience, and flexibility for use.
AES block cipher receives a 128-bit plaintext, and then creates a 128-bit ciphertext under
the control of the 128-bit key. It is an alternative—exchange network design with 16 rounds
(Round) iterative process and a plaintext is changed to a ciphertext after the process.
A round of AES is composed of the following four steps:

1. SubByte
2. ShiftRows
3. MixColumns
4. AddRoundKey

TABLE 4.6
Comparison between Symmetric Encryption and Asymmetric Encryption
Asymmetric Encryption Symmetric Encryption
Performance The algorithm is complex and The algorithm is simple and effective
the encryption speed is low. with high encryption speed
Key management Only private key is confidential Secret key must be established and
protected confidentially by both sides
The confidentiality period of The confidentiality period of secret
public/private key is long key is very short
84 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

IV shift register IV shift register IV shift register

...
Key Encryption Key Encryption Key Encryption

Use the left j bits Use the left j bits Use the left j bits

XOR XOR XOR

j bits plaintexts j bits plaintexts j bits plaintexts

j bits ciphertexts j bits ciphertexts j bits ciphertexts

FIGURE 4.5
CFB mode flow diagram.

The 128-bit round key used in each round is generated by the secret key through the pro-
cess that is referred to as a key schedule.
CFB mode allows the length of encrypted data arbitrary, so it is suitable for character-
oriented applications. Every j bits are processed in CFB mode and specific encryption
processes are only needed to repeat the following steps, until all plaintext units are
encrypted15:
The first step is to use the initialization vector IV and key to create encryption vector.
The second step is to make XOR (exclusive or operation) between j bits plaintexts and
encrypted vector to get the j bits ciphertexts.
The third step is to move the initialization vector j bits to the left, and to fill j bits ciphertext.
CFB128 is adopted dealing with 128-bit at one time and the specific process is shown in
Figure 4.5.

4.3.4.2  Message Authentication Code

MAC is very important in ACARS information security policy because of both the data
integrity and message authentication services, but also because the defense of DoS cannot
succeed without it.16
The purpose of MAC is to ensure that both of the two sides sharing a secret key (or more)
in communication have the ability to check whether the transmitted information is modi-
fied or not. Further, it can also guarantee the identity authentication between two or more
sides in communication.
The news received and secret key are the input of the MAC algorithm that produces a
MAC tag with fixed size to finish this work. News and the tag will be transmitted to the
receiver, to recalculate the tag and then to compare the transmitted tag with the recalcu-
lated tag. If they are same, the news is almost certainly authentic and correct. Otherwise,
Security of VHF Data Link in ATM 85

the news is not right, and should be discarded, or the connection needs to be given up,
because it may have been tampered with. The attacker who wants to forge or tamper with
the message needs to break the MAC function and it is obviously not an easy thing.
HMAC is selected as the MAC algorithm in ACARS security scheme. HMAC, uses one-
way hash function on cryptography and transforms it to MAC algorithm.
It needs a hash function used to encrypt17 (marked as H, may be MD5 or SHA-1 or SHA-
256 is selected in this chapter) and a key to define the HMAC. We use B to represent the
number of data block bytes (the word length of segmentation data block of the given hash
functions B = 64), and use L to represent the number of output data bytes of hash function
(L = 16 in MD5, L = 20 in SHA-1, and L = 32 in SHA-256). The length of the authentication
key can be any positive integer that is not bigger than the word length of the data block. If
the length of the key is bigger than B, the key would be processsed by hash function H. The
output of hash function H is a L-length string which can be used in the HMAC. In general,
the recommended minimum key length K is L bytes. In general, the recommended mini-
mum key length K is L bytes.
We define two fixed but different strings, ipad and opad: (i and o represent internal and
external)

ipad = the byte 0 × 36, to repeat B times

opad = the byte 0 × 5C, to repeat B times
To calculate HMAC of text
HMAC=H(K XOR opad, H(K XOR ipad, text))

As the following steps:

1. To create a B-length string by adding 0 after the key K (for example, if the length
of K is 20 bytes, B = 64 bytes, then 44 zero bytes 0 x00 should be added after the K)
2. To make XOR between B-length string generated by the first step and ipad
3. To fill the data flow text into the result string of step (2)
4. To process the data flow generated in step (3) by H
5. To make XOR between B-length string generated by the first step and opad
6. To fill the result of step (4) into the result of step (5)
7. To process data flow generated in step (6) by H, and the output is the final result

A structure diagram is shown in Figure 4.6.

4.3.4.3  Digital Signature Scheme

The authentication of a message in ACARS requires authentication between peer entities
to guarantee reliability. For example, the plane needs authentication from the ground sta-
tion and the ground station also needs authentication of the information acquired from
the plane. The integrity should guarantee the content integrity and sequence integrity of
the ACARS message.18
By using public key encryption technology, we can proceed with the message authenti-
cation and guarantee data integrity. The public key encryption technology includes many
applications, such as the RSA algorithm based on the integer factorization problem, the data
signature algorithm (DSA) based on the discrete logarithm problem, the Diffie–Hellman
86 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Key Message

Process ipad
the key

XOR Hash

opad

XOR Hash

Mark

FIGURE 4.6
Structure diagram of HMAC.

(DH) algorithm, the ECDSA algorithm based on the elliptical curve problem, the ECDH
algorithm,19 and so on. RSA algorithm, DSA algorithm and ESDSA algorithm can provide
the message authentication service perfectly. The method is that the sender uses the pri-
vate key to encrypt, and then the receiver uses the public key to decrypt. By this way, as
the key is private, the receiver can identify the sender and guarantee the integrity of the
message. The security of public key encryption technology depends on the length of the
key. For example, the elliptical curve algorithm has a shorter key in the same security
rank compared with other algorithms, so it is better adapted to the limited resources of
ACARS. Furthermore, the purpose of the design of the data signature is to carry out iden-
tity authentication and to avoid deception and denial. Hence, we can choose the ECDSA
algorithm.20
In the ACARS information security solution, there are two ways to establish a secure
connection: shared key and public/private key, and the latter uses the ECDSA as the ini-
tial authentication method when it establishes connections. However, in the process of
the data transmission, the system mainly uses MAC to guarantee message integrity and
authentication.

4.3.4.4  Key Management

All kinds of security services are based on different encryption algorithms, and the
security of the algorithm relies on the security of the key, so key management involves
the security and reliability of all security services. Key management involves many
Security of VHF Data Link in ATM 87

aspects including key generation, distribution, utilization, updating, storage, backup,

and destruction.21
There are two kinds of keys in aircraft and ground communication applications: one is
the asymmetric key used for digital signature or certification and the other is the symmet-
ric key used for link encryption and integrity services.

4.3.4.4.1  PKI-Based Key Management

Elliptic curve cryptographic (ECC) algorithms are employed to perform secure session
initiation and key establishment between communicating peer AMS aircraft and ground
entities in the case that AMS is implemented using public/private keys. Public key cryp-
tography uses a pair of asymmetric keys, which are related mathematically, and one-way
functions, and the important characteristic of the keys is that it is easy to compute using
one of the keys but very difficult to solve without knowing the related key. One of the keys
must be kept private and protected from compromise to ensure security, while another
one may be distributed publicly. Although one key is public, it is very difficult to derive the
private key by knowing just the public key that is made by the mathematical properties of
cryptographic algorithms such as ECC.12
PKI provides life-cycle management of the public/private keys to ensure secure sessions
for AMS entities. CA, a trusted PKI entity, issues public key certificates to AMS entities.
During the issuance, CA digitally signs the content of the public key certificate and the
public key certificate binds the public portion of the key pair with the identity of the AMS
entity that uses the key pair.
Since there is no sensitive cryptographic information in the public key certificate and the
public key is protected by the CA digital signature from modification, the CA may publish
the public key certificate via an unprotected channel. Before using public key certificates,
AMS entities must verify the CA’s digital signature to ensure that the certificate has been
certified and has not been modified during the storage or distribution.
Since the public key certificates in asymmetric cryptography are permitted to be
distributed freely, PKI-based key management is particularly suited for the following
situations:

• It is practical and feasible to preplace and/or maintain the pair-wise shared-secret

keys.
• The quantity of entities increases over time. So scalability is important.
• Security services must interoperate across multiple, heterogeneous organizations.
For example, AMS DSP-based security is that one or more service providers offer
AMS security services to multiple airline organizations.
• There is a need to know the identity of the entity with which secure communica-
tion is being established, for example, an aircraft crew needs to authenticate the
identity of a ground entity and vice versa.

4.3.4.4.2  Shared Secret Key Management

Symmetric algorithms are employed to perform secure session initiation and key estab-
lishment between communicating peer AMS aircraft and ground entities when AMS is
implemented using a shared secret key. Each AMS entity is required to share a secret key
in the symmetric algorithms. Also, each AMS entity must protect the shared secret key
from compromise to ensure security. If the key is compromised for one entity, then it is
equally compromised for all entities that share the same key.12
88 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

A key management infrastructure (KMI) provides the necessary life-cycle management

of the shared secret keys to ensure secure session for AMS entities. The KMI is respon-
sible for generation and distribution of a shared secret key to each pair of AMS entities.
The shared secret key is distributed to each AMS entity through a secure communication
channel independent of the ACARS datalink.
Shared secret–based key management is particularly suited for the situations as follows:

• The quantity of entities is small and generally does not change over time. So scal-
ability is not important.
• Security services must be within an organization for interoperability among enti-
ties. For example, AMS end-to-end security is that one airline manages all of the
aircraft entities and ground entities that must communicate securely.
• It is practical and feasible to preplace and/or maintain the pair-wise shared-secret
keys.
• There is no need to know the identity of the entity with which secure communica-
tion is being established.

4.3.4.5  Payload Encoding

ACARS is a character-oriented communication system, but almost all encryption algo-
rithms are bit-oriented until now. Hence, the ACARS message should be converted to
bit stream before it is encrypted by encryption algorithm. However, the characters used
in the ACARS message are all noncontrol characters in the ISO 5 character set, and they
rarely use lowercase letters. Meanwhile, the information coding before transmission
will increase the length of the message. As a result, it is really necessary to load coding
for the ACARS message.
The ACARS message should first be converted from character to bit using an encryption
algorithm. ACARS uses the ASCII character set apart from the control character that does
not appear in the user data stream;  we have organized the most frequently used 64 char-
acters as the coding in Table 4.7 after analyzing a large number of ACARS messages. Based
on the table, an 8-bit character can be converted to 6-bits, and then they are connected as
bit stream to decrease the effective load by 25%.
As to the characters apart from the coding table, we delete the most significant bit 0 to
reduce the 8-bit character to a 7-bit character stream, and the effective load can be decreased
by 12.5%. The length of the bit stream will be an integral multiple of eight no matter which
coding method is selected. Otherwise, it needs to pad with padding bit by bit.
In the recipient, the original ACARS message data are available after the bit stream being
decoded is anticoded based on the coding table.
This coding way not only converts the ACARS message to bit stream for data encryp-
tion, but also allows for information coding later to be more effective and practical.

4.3.4.6  Information Coding

All ACARS security services, including data encryption and the ACARS message
with MAC, are required to undergo information coding before they are transferred by
the ACARS network. Information coding also uses a 64-bit character coding table, and
the information coding table is basically the same as the load coding table, apart from the
character LF (0 × 0A) and CR (0 × 0D) that are replaced by “[” (0 × 5B) and “]” (0 × 5D).
TABLE 4.7
Coding Table of 6 Bits Effective Load
Character Character Character Character
6 Bits 6 Bits 6 Bits 6 Bits
ASCII HEX Coding ASCII HEX Coding ASCII HEX Coding ASCII HEX Coding
SP 0 × 20 000000 0 0 × 30 010000 @ 0 × 40 100000 P 0 × 50 110000
Security of VHF Data Link in ATM

! 0 × 21 000001 1 0 × 31 010001 A 0 × 41 100001 Q 0 × 51 110001

“ 0 × 22 000010 2 0 × 32 010010 B 0 × 42 100010 R 0 × 52 110010
# 0 × 23 000011 3 0 × 33 010011 C 0 × 43 100011 S 0 × 53 110011
$ 0 × 24 000100 4 0 × 34 010100 D 0 × 44 100100 T 0 × 54 110100
% 0 × 25 000101 5 0 × 35 010101 E 0 × 45 100101 U 0 × 55 110101
& 0 × 26 000110 6 0 × 36 010110 F 0 × 46 100110 V 0 × 56 110110
‘ 0 × 27 000111 7 0 × 37 010111 G 0 × 47 100111 W 0 × 57 110111
( 0 × 28 001000 8 0 × 38 011000 H 0 × 48 101000 X 0 × 58 111000
) 0 × 29 001001 9 0 × 39 011001 I 0 × 49 101001 Y 0 × 59 111001
* 0 × 2A 001010 : 0 × 3A 011010 J 0 × 4A 101010 Z 0 × 5A 111010
+ 0 × 2B 001011 ; 0 × 3B 011011 K 0 × 4B 101011 LF 0 × 0A 111011
, 0 × 2C 001100 < 0 × 3C 011100 L 0 × 4C 101100 \ 0 × 5C 111100
– 0 × 2D 001101 = 0 × 3D 011101 M 0 × 4D 101101 CR 0 × 0D 111101
. 0 × 2E 001110 > 0 × 3E 011110 N 0 × 4E 101110 ˆ 0 × 5E 111110
/ 0 × 2F 001111 ? 0 × 3F 011111 O 0 × 4F 101111 | 0 × 7C 111111
89
90 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Read-in bit stream

data

No
Remaining bits
>=8

Yes

Read-in 8 bits, then to get the corresponding

characters according to the ASCII

Yes
No
Transportable?

Yes

No
In 64-bit code
table?

Yes

Encoding by Encoding by 6 bits Encoding by 6 bits (zero

original 8 bits character pad)

FIGURE 4.7
Information coding process.

First, we convert the message to bit stream. Eight bits are read from high bit, and it is
judged whether they are characters that could be transferred. If the answer is no, then we
code the first 6 bits to an 8-bit character based on the information table. If the answer is yes,
we continue to judge whether the messages are in the information table. If the answer is
yes, we code the first 6 bits to an 8-bit character based on the information coding table. If
the answer is no, we code the original 8 bit. According to such steps over and over again,
until the remaining number of bits are equal to or less than 6 bits, we code the message
based on the information in the coding table (zero is padded while the length of the mes-
sage is shorter than 6 bits). Finally, we get the result. The process is shown in Figure 4.7.
After information coding, the final message consists of all transportable ACARS charac-
ters and the result could be transferred by the ACARS network.

4.3.4.7  Data Compression

The purpose of data coding and data compression is to reduce the effect of the security
load as much as possible. The AMS standard includes coding and compression for the
effective load, which lowers the length of the AMS information. Besides, the data com-
pression program and the AMS security mechanism produce arbitrary binary bit stream
and an information coding mechanism offers a harmless method to convert the bit-ori-
ented information stream to a character-oriented information stream so as to transfer to
the ACARS network.
Security of VHF Data Link in ATM 91

AMS offers the mechanism of a compression algorithm. Both sides of the communica-
tion offer a compression algorithm supported by themselves, and an optimal algorithm is
finally chosen by negotiation. The compression rate is determined by the original informa-
tion. Generally speaking, compared to shorter and more random information, the longer
information with a higher repetition rate has a higher information compression rate. AMS
offers two default compression algorithms, DMC and DEFLATE.

4.3.5 Conclusion
This chapter analyzes some main security threats for the ACARS datalink and offers
detailed solutions to a variety of threats. Meanwhile, it introduces varieties of algorithms
used in the solution, such as encryption, authentication, key exchange, and coding and
decoding.

References
1. Z. Xuebing and M. Yuwen. VHF air/ground datalink system and ARINC 618 protocol. Aviation
Maintenance and Engineering, 2002, 2(1):41–43.
2. ARINC. Datalink security part 1–ACARS message security, ARINC specification 823P1.
Annapolis, MD: Aeronautical Radio, Inc., December 10, 2007.
3. X. Xiaogang. The Research and Application of Data-Link Technology in Tower Control. Beijing, China:
Beihang University, 2004.
4. H. Junxiang. Research on ACARS data process. Journal of Civil Aviation University of China, 2007,
25(1):1–3.
5. X. Wenhui. Research on the realization technology of ACARS. Avionics Technology, 1995,
26(2):25–28.
6. ARINC. ARINC specification 618-5: Air/ground character-oriented protocol specification.
Annapolis, MD: Aeronautical Radio, Inc., 2000.
7. G. Hong, J. Jun, and X. Wenyan. Parameter identifying of the ACARS message. Avionics
Technology, 2006, 37(4):6–11.
8. ARINC. ARINC specification 620-4: Data link ground system standard and interface—
Specification (DGSS/IS). Annapolis, MD: Aeronautical Radio, Inc., 1999.
9. H. Jia. Air/ground data link system and ARINC 620 protocol. Aviation Maintenance and
Engineering, 2004, 5(1): 33–35.
10. W. Xiaolin, Z. Xuejun, and H. Jia. Security communications in ACARS data link. Avionics
Technology, 2003, 34(zl):95–100.
11. C. Hao, H. Yisheng, J. Tao, X. Yongchun, and W. Yunbing. Status and development of the data
link security techniques. Electronic Warfare, 2006, 111(6):37–41.
12. ARINC. Datalink security part 1—Key management, ARINC specification 823P2. Annapolis,
MA, ARINC, March 10, 2008.
13. S. Tao, M. Hongguang, and X. Wentong. Research and application on network data encryption.
Computer Engineering and Application, 2002, 38(19):156–158.
14. National Institute of Standards and Technology. Advanced encryption standard, draft fed-
eral information processing standard. Gaithersburg, MD: National Institute of Standards and
Technology, 2001.
15. A. Kahate. Cryptography and Network Security. Beijing, China: Tsinghua University Press.
16. H. Krawcyzk, M. Bellare, and R. Canetti. HMAC: Keyed hashing for message authentication.
Internet Engineering Task Force, Internet RFC 2104, 1997.
92 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

17. W. Hongxia, L. Saiqun. A mechanism for message authentication using HMAC-SHA1. Journal
of Shanxi Teacher’s University (Natural Science Edition), 2005, 19(1):30–33.
18. L. Zhigui, Y. Lichun, P. Jie et al. The system of digital signature authentication based on PKI.
Application and Research of Computers, 2004, 9:58–160.
19. V. Patel and T. McParland, Public key infrastructure for air traffic management systems. 20th
Digital Aviation Systems Conference, Daytona Beach, FL, October 2001.
20. American National Standards Institute. Public key cryptography for the financial services
industry—The elliptic curve digital signature algorithm (ECDSA), ANSI X9.62. American
National Standards Institute, 1998.
21. Y. Zhijun. Distribution and management of online security and key. Journal of Taiyuan Teachers
College (Natural Science Edition), 2004, 3(3): 22–24.
5
VDL2 Key Technology and Simulation

Gao Lin

CONTENTS
5.1 VDL General Introduction................................................................................................... 94
5.1.1 Transition from ACARS to the VDL....................................................................... 94
5.1.2 VDL Modes................................................................................................................ 95
5.1.2.1 VDL2 Character.......................................................................................... 95
5.1.2.2 VDL3 Character.......................................................................................... 97
5.1.2.3 VDL4 Character.......................................................................................... 97
5.1.3 VDL Modes Comparison......................................................................................... 99
5.2 VDL Protocol....................................................................................................................... 100
5.2.1 Stacks Structure...................................................................................................... 101
5.2.2 Physical Layer.......................................................................................................... 103
5.2.2.1 Architecture.............................................................................................. 103
5.2.2.2 Functionality............................................................................................. 103
5.2.2.3 Physical Frame Structure........................................................................ 104
5.2.3 Data Link Layer....................................................................................................... 105
5.2.3.1 Media Access Control.............................................................................. 105
5.2.3.2 Data Link Service..................................................................................... 106
5.2.3.3 VDL Management Entity........................................................................ 107
5.2.4 Subnetwork Layer................................................................................................... 108
5.2.4.1 Architecture.............................................................................................. 109
5.2.4.2 Functionality............................................................................................. 109
5.3 Key Technology................................................................................................................... 109
5.3.1 Physical Layer Key Technology............................................................................ 109
5.3.1.1 D8PSK Modulation.................................................................................. 109
5.3.1.2 RS Channel Encoding............................................................................. 111
5.3.2 Data Link Layer Key Technology......................................................................... 112
5.3.2.1 P-CSMA..................................................................................................... 112
5.3.2.2 From HDLC to AVLC.............................................................................. 114
5.3.2.3 Handoff...................................................................................................... 116
5.3.3 Subnetwork Link and Connection....................................................................... 118
5.3.3.1 Explicit Subnetwork Connection........................................................... 118
5.3.3.2 Expedited Subnetwork Connection...................................................... 119

93
94 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

5.4 Modeling and Simulation.................................................................................................. 120

5.4.1 VDL2 Modeling....................................................................................................... 120
5.4.1.1 Physical Link Layer Modeling............................................................... 121
5.4.1.2 Data Link Layer Modeling...................................................................... 122
5.4.1.3 Packet Layer Modeling............................................................................ 125
5.4.2 Handoff Simulation................................................................................................ 128
5.4.2.1 SINR-Based Handoff............................................................................... 128
5.4.2.2 Experiment Setup.................................................................................... 130
5.4.2.3 Simulation Results................................................................................... 132
5.5 Conclusion........................................................................................................................... 135
References...................................................................................................................................... 135

5.1  VDL General Introduction

The very high-frequency digital link (VDL) is an air–ground subnetwork that may be
applied in supporting data communication across aeronautical telecommunication net-
work (ATN) aircraft-based application processes and the peer ground-based processes.
Presently, the VDL has been recommended as the major method of air traffic management
(ATM) in the transition to the future ATN and has been deployed in many areas such as
central Europe and North America.

5.1.1  Transition from ACARS to the VDL

Before the advent of the data link system, all communications between the ground staff
and flight crew were only through voice by means of very high-frequency or high-
frequency voice radio communication. To reduce the workload and improve data integ-
rity, Aircraft Communication Addressing and Reporting System (ACARS) was proposed
as the first aeronautical air–ground data link in the 1970s. ACARS was first researched and
developed by ARINC. Afterward, ARINC and SITA have become the two main providers
of the ACARS system in the world. For a long time, it has been a very popular aeronauti-
cal data link and has become the dominant air–ground communication mode worldwide.1
ACARS works in very high frequency (VHF) with 25 kHz channel band. The modulation
scheme of the system is amplitude modulation minimum shift keying (AM-MSK). It has no
forward error correction (FEC). The channel access scheme is nonpersistent carrier sense
multiple access (CSMA). It is character oriented and can communicate only at 2.4 kbps.
With the development of air service, ACARS has become unable to satisfy the high-
capacity need of aviation communication. VDL technology emerges as the time requires.
It can arrive at far more excellent performance than ACARS. For example, the VDL2, as a
typical VDL mode, has a 13 times higher transmission rate than ACARS.
The VDL2 is a bit-oriented packet communication system with 25 kHz channel separa-
tion. The modulation scheme of the system is differential 8 phase-shift keying (D8PSK).
The VDL2 has a FEC function with Reed–Solomon (RS) code. The channel access scheme
is p-persistent CSMA. It can get a 3.15 kbps transmission bit rate.
A performance comparison between the VDL2 and ACARS has been made by simula-
tion.2 In the simulation, the assumptions were made that the message data length and data
generation interval, respectively, followed uniform distribution and Pareto distribution
VDL2 Key Technology and Simulation 95

according to flight data analysis. Also, the load condition approximated the actual ACARS
data ratio (Table 5.1). The conclusions are as follows:

1. When the data traffic load generated from an aircraft equaled the load directed to the
aircraft, the VDL2 was able to process 4.6 times more congested load than ACARS.
2. When the data traffic load generated from an aircraft was five times higher than
the load directed to the aircraft, the VDL2 was able to process 8.8 times more con-
gested load than ACARS.

Besides the VDL2, modes 3 and 4 also have their advantages over ACARC in performance.
Compared to the VDL, there are three defaults in ACARS, which restrict the system per-
formance. First, AM mode determines that the frequency utilization rate is relatively low,
and the system is susceptible to radio-frequency (RF) interference. Second, the communi-
cation technology of the ACARS data link, as the character-oriented system, is less flexible
than bit-oriented one. Third, not like the VDL, ACARS is not compatible with open sys-
tems interconnection (OSI), which is adopted by the ATN.
Under the growth in air traffic and the development of the ATM services, the VDL is
attracting more attention of civil aviation organizations and institutions in major coun-
tries and regions of the world. The VDL concerned test and deployment works are being
carried out systematically. The aeronautical data link is nonexperiencing a transition from
ACARS to the next-generation data link—the VDL.

5.1.2  VDL Modes

As yet, the ICAO has approved four VHF modes as the ATN-supporting data link and has
written them in Annex 10.3 Among them, the VDL1 is the intermediate path from the exist-
ing ACARS to the VHF data link modes and has less practical value. Accordingly, only
three modes are described and discussed in the subsequent sections.

5.1.2.1  VDL2 Character

The VDL2 is the air–ground data link designated by ICAO aviation mobile communica-
tions experts group (Aeronautical Mobile Communications panel [AMCP]) in 1997. There
are many countries involved in its development and construction. Europe is the main
VDL2 supporter and implementer. They began to test the feasibility of using the VDL2
data link in 2001 and have carried out the Link2000+ project relying on Eurocontrol. Japan
is another VDL2 supporter and has built two ground stations (GSs) by early 2004.4
In the VDL2, the channel spacing is 25 kHz, and the modulation scheme is D8PSK. In
the data link layer, the VDL2 applies CSMA in the medium access control (MAC) sublayer.
In the data link service (DLS) sublayer, it specifies the aviation VHF link control (AVLC)
protocol, which is derived and developed from high-level data link control (HDLC) proto-
col. The VDL2 transmission rate is 31.5 kbps. (For details of the protocol and key technol-
ogy of the VDL2, refer to the next two sections in the chapter.)
The VDL2 can provide air–ground communication and uplink broadcast service. It can
provide Aeronautical Operational Communication, as well as be used for CPDLC, ATIS,
ATC, etc., applications. However, the VDL2 does not support voice so far. Also, it does not
provide priority function, which increases the information delay in the high offered load
circumstance and makes it unsuitable for emergency use.
 96

TABLE 5.1
Comparison between ACARS and the VDL2
Channel Information Information Error Supported Transmission Bit
System Modulation Space (kHz) Unit Type Correction Access Scheme Protocol Rate (kbps)
ACARS AM-MSK 25 Character Date No Nonpersistent CSMA ACARS 2.4
VDL2 D8PSK 25 Bit Date Yes p-persistent CSMA ATN ACARS 31.5
Aeronautical Telecommunications Network: Advances, Challenges, and Modeling
VDL2 Key Technology and Simulation 97

5.1.2.2  VDL3 Character

The VDL3 is the data link that FAA energetically pushes for implementation in the
United States. FAA intends to use multimode digital radio to gradually replace the exist-
ing ground VHF/UHF stations in the United States.
The VDL3 can transmit both data and voice. As with mode 2, the VDL3 applies 25 kHz
channel spacing and specifies a D8PSK modulation scheme operating at 31.5  kbps. The
voice coding transmission rate is 4.8 kbps (normal voice) or 4 kbps (truncated voice).
In the VDL3, the MAC sublayer applies time division multiple access (TDMA) in which
slot usage for data transmission is controlled by the GS. Several configurations are applied
with the standard four-slot per frame and long-range three-slot per frame timing struc-
tures to accommodate different amounts of voice and data traffic. In the standard range
configurations, four 30 ms time slots, denoted as A, B, C, and D, make up a MAC frame.
A MAC cycle consists of two MAC frames, even and odd, with a duration of 240 ms. Time
slots are subdivided into Logical Burst Access Channels (LBACs) to provide management
capabilities in the voice and data slots.
The VDL3 supports four levels of priority. Based on priority, the ground decides when
to grant access to the link. Reservation responses are sent in the uplink management (M)
burst in slot A of the odd frame. If the ground is unable to immediately grant the request
when it is received, it responds with a request acknowledgment (RACK). Otherwise, it
will indicate which slot in the next MAC cycle may be used to begin the transfer, as well
as whether slot D may be used or if it is reserved for voice. Once access has been granted
to the link, a station will transmit starting in the indicated slot and apply consecutive data
slots until the transmission is completed. Transmissions are limited to a maximum of 15
consecutive slots. In the event that a voice reservation is also granted, only slots B and C
will be used for data transmission.
DLE uses an acknowledged connectionless protocol (ACLP) that adds only 6  bytes
of header information to each network frame. Upon sending a frame, DLE waits for an
acknowledgment before sending the next frame. The MAC sublayer notifies the DLE
when an ACK is not received at the expected time. Frame grouping is allowed at the
DLS sublayer, in which frames of the same priority and destination may be sent together.
Optionally, a frame of lower priority may also be included in the group as long as it does
not require additional slots to transmit. An ACK will acknowledge all frames in the group.
If the T1 timer expires before an acknowledgment is received and retransmission occurs,
and a higher priority packet is queued, the retransmission of the unacknowledged group
will be delayed.
Downlink acknowledgments created by the DLE are automatically converted into an
ACK burst and scheduled for transmission in the M burst LBAC after the last data segment
of one MAC cycle is sent. The GS recreates the DLEACK from the ACK burst and passes
it to the DLE. Uplink acknowledgments are sent in normal data bursts, but with an expe-
dited priority that takes precedence over any ungranted data transmission.
The VDL3 can support voice transmission and can provide air–ground communication
and uplink broadcast, etc., ATN services. At the same time, it is fully compatible with the
ATN in data transmission.

5.1.2.3  VDL4 Character

The VDL4 is the scheme proposed by Sweden. With the global navigation satellite system
(GNSS) information timing and the bit-oriented protocol, the VDL4 can support multiple
98 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

communication services for the ATN and even can provide more than the ATN services,
such as ADS-B. At present, Eurocontrol is committed to the development of the VDL4 for
the communication and surveillance services, including the system standard, frequency,
and structure.
As with modes 2 and 3, the VDL4 channel spacing is 25 kHz, but the physical layer
uses a Gaussian-filtered frequency shift keying (GFSK) modulation operating at 19.2 kbps.
MAC uses a self-organizing time division multiple access (STDMA) mechanism. Time is
segmented into 13.3 ms timeslots, with 75 slots per second. A superframe lasts for 60 s and
contains 4500 slots. Aircraft timing is primarily based on GNSS, but alternatively, timing
can be derived from GSs or other sources, including other aircraft. This allows the system
to continue operating even when communication with the primary source fails.
The MAC burst for mode 4 requires more information to be contained within it than
the other modes. Mode 4 bursts contain a start and an end flag, reservation information,
source address, message, and a cyclic redundancy check (CRC). The size of the burst
header depends on the type of message being transmitted and the type of reservation
being placed.
The VDL4 receivers are intended to be multichannel. Two global signaling channels
are defined for mode 4. Aircraft are expected to transmit position information, such as in
ADS-B messages, alternately on these two channels. In areas where traffic levels are high,
the aircraft can be directed to local channels for use.
The VDL4 has two operational modes: autonomous reporting and directed reporting.
When operating under autonomous reporting, each aircraft chooses its own slots for trans-
mission, whereas in directed reporting, the GS determines which slots aircraft may use.
Slot selection is handled by the VSS. When a reservation is to be placed, the VSS iden-
tifies the first Q4 available slots in the range specified by the application that meet the
length requirement. In the event that less than Q4 available slots are found in the speci-
fied range, a mechanism exists to select previously reserved frames from distant aircraft
communicating with distant stations. This Robin Hood protocol allows for slot reuse if
certain conditions are met. From the available list of slots, the VSS randomly selects one
for use.
Several reservation mechanisms exist for the VDL4, such as periodic broadcast, incre-
mental broadcast, unicast request, information transfer request, directed request, super-
frame block, and second block reservation protocols. The unicast reservation allows for
a one-way transfer. A station can reserve a block of slots on behalf of another station.
The information transfer protocol, however, is a two-way communication. A station
reserves slots for another station to use, and it also reserves a single slot for itself to send
an acknowledgment.
The unicast reservation utilizes a VSS-based retransmission mechanism. If a response
is not received from the peer in the slot reserved for the transfer, the retransmission pro-
cedures will be invoked. Retransmitted frames access the link by random access. The
retransmission delay is dynamically calculated based on slot utilization and the number
of retransmission attempts. The value of Q5 is randomly set, according to the standards,
between a minimum and a maximum value, where the maximum is determined by the
number of retransmissions and the minimum by default is 0.
Information transfers do not use the VSS for retransmission; the retransmission is the
responsibility of the DLS sublayer. Upon indication that an information transfer is not
successful, a DLE will restart the transmission process. The DLE uses a negotiated setup
connection-oriented protocol (NSCOP) for air-to-ground communications and a zero
VDL2 Key Technology and Simulation 99

overhead connection-oriented protocol (ZOCOP) for air-to-air communications. The DLS

supports 16 levels of priority, enough to map each of the 15 levels of priority specified by
the ATN. The larger burst format at the MAC sublayer allows for a small DLE frame header
to be used. For data packets, the header is only 2 bytes in length.
The DLS will use either the short or long transmission procedures depending on the
length of the frame. In the short procedure, the DLS sends a data frame to the VSS with
instructions to include a unicast reservation for a single slot with the transmission. The
VSS will place a reservation for the data frame in an existing transmission, if possible;
otherwise, it will specify that the frame be transmitted by random access. When the
frame is received, an acknowledgment is created by the peer DLS and is transmitted
back to the source. The acknowledgment will be placed in the slot reserved by the uni-
cast reservation.
For a long transmission, the DLS issues a request-to-send (RTS) frame that includes the
length of the data transmission. The RTS is sent similarly to the data frame of the short
transmission, with a unicast reservation for a single slot for the peer. The peer, upon receiv-
ing an RTS, responds with a clear-to-send (CTS). The CTS is sent in the frame reserved for
it and includes an information transfer reservation for the data length specified in the RTS.
The information transfer reservation reserves slots for the data frame as well as a slot for
the acknowledgment. The DLS will send the frame when it receives the CTS, and the peer
will respond with an acknowledgment.
DLE allows linking of transmissions such that data and RTS frames can be combined in
a single transmission. In these cases, the ACK and CTS responses will also be combined.
When transmissions are linked in this way, all transmissions after the initial RTS will be
by reserved access.
The VDL4 is fully compatible with the ATN network. It can realize the air–ground com-
munication, uplink and downlink broadcast, air-to-air broadcast, etc., ATN services. It also
can provide more than the ATN services, such as ADS-B and air-to-air communication. As
its limitation, the VDL4 does not support voice.

5.1.3  VDL Modes Comparison

We can compare the modes from the aspects of system performance, support service, tech-
nology compatibility, etc.

1. The VDL2 has the highest capacity, which can be valued by the supported aircraft
(AC) number under the delay required by corresponding regulation. The VDL
MASPS states that the 95th percentile subnetwork delay should be no more than
3.5 s for mode 2. For mode 3, the maximum 95th percentile subnetwork delay for
high-priority traffic of 192 bits or less is 1 s, and the maximum 99.9th percentile
subnetwork delay is 5 s, which is shorter and stricter than the VDL2 due to the pri-
ority support functionality.5 Since mode 4 has only been standardized for use in
the ATN for surveillance and not as a communications data link, a maximum limit
does not exist and is substituted by the mode 3 definition. Under the delay limits,
simulations determined that Mode 2 is capable of bearing 3.3  kbps load, which
is equivalent to the load from approximately 130 aircraft. Compared to mode 2,
modes 3 and 4 have far less load with only mode 3 supporting 0.5 kbps and mode
4 supporting 0.8 kbps, which are equivalent to the load from approximately 20 and
30 aircraft, respectively.6
100 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 5.2
VDL Modes’ Comparison
Media Voice Estimated Max Transmission
Mode Modulation Access Transfer Control Priority Support Load (kbps) Bit Rate (kbps)
VDL2 D8PSK CSMA AVLC No No 3.3 31.5
VDL3 D8PSK TDMA ACLP 4 levels Yes 0.5 31.5
VDL4 GFSK STDMA NSCOP/ZOCOP 15 levels No 0.8 19.2

2. In the terminal and en route domain simulations, the VDL2 delay is on the whole
better than the VDL3 and almost equivalent to the VDL4. The experiments were
under the same traffic model with typical services and message rates. With the
default parameters regulated by the corresponding protocols, the three modes
result in similar average uplink delays. In the downlink, modes 2 and 4 can pro-
vide lower average 95th percentile delays than mode 3. With the optimized param-
eters, mode 2 has 99.9th percentile delays rivaling those of modes 3 and 4 in the en
route domain and is better than modes 3 and 4 in the terminal domain.
3. In service aspect, modes 2 and 4 cannot support voice, while mode 3 can support
both data and voice service. Mode 4 can supply diverse date service for both ATN
and non-ATN, for example, the VDL4 can support ADS-B and air-to-air communi-
cation, which are the services more than the ATN. On the other hand, modes 3 and
4 support priority and can be used for safety-of-flight messages. Because of CSMA
implementation, the VDL2 cannot provide prioritized link access and is intended
to communicate noncritical information.
4. Different from modes 3 and 4, the VDL2 not only support the ATN, but also
ACARS. The VDL2 can support messages in ACARS format and can deal with
ACARS protocol by ACARS over AVLC technology. The VDL2 also has more simi-
larity in implementation with ACARS than other modes, for example, the two
systems implement CSMA for media access. Due to these reasons mentioned, the
VDL2 has the best compatibility with ACARS, which has the largest data link
share in the current aviation industry.

From the VDL mode comparisons, we can know that the VDL2 has the best performance
as a whole (Table 5.2). It can achieve the highest capacity and almost has the shortest delay.
Besides, the VDL2 system has the best compatibility with ACARS among the VDLS, which
can benefit and smooth the transition from the current system to the future ATN. At the same
time, the deficiency of the VDL2 service (e.g., it does not support voice and priority) can be
made up by using another communication system, such as aeronautical mobile satellite sys-
tem. In summary, the VDL2 is the most hopeful VHF data link to succeed ACARS.

5.2  VDL Protocol

As an ATN-compatible data link, the VDL2 has been approved by the ICAO. The VDL2-
related protocol is stated in Standards and Recommended Practices (SARPs), which were
developed by the AMCP7 and introduced in the ICAO Annex 10.8
VDL2 Key Technology and Simulation 101

5.2.1  Stacks Structure

The international aviation community is expected to adhere to the separation of commu-
nication functions as specified in the OSI reference model developed by the International
Organization for Standards (ISO). The OSI reference model permits the development of
open communications protocols as a layered architecture comprising seven functional
separate layers.
As an entire network, the ATN conforms to the OSI reference model and has all of its
seven layers, including low-level three layers and high-level four layers. The VDL2, which
is the ATN communication subnet and is fully compatible with the ATN, also conforms
to the OSI reference model but only has the low level. It constitutes the first step toward a
fully OSI-compatible protocol stack (Figure 5.1).
The ATN can be divided into ground subnetwork, avionics subnetwork, and air–ground
subnetwork. Among them, ground subnetwork and avionics subnetwork belong to fixed
data links, which, respectively, realize in aircraft (AC) and ground; air–ground subnetwork
belongs to mobile data link. The subnetworks can be regarded as being made up of end
system (ES) and intermediate system (IS), which are two elemental systems abstracted
from real equipment. ES can be defined as the computer system that executes end-user
application programs and communicates to IS or possibly to other ESs. ES can exist in both
avionics subnetwork (e.g., display control system of aircraft cockpit) and ground subnet-
work (e.g., the ATM system of GS). IS can be defined as the computer system that relays
and routes network protocol data unit (NPDU) to other ISs or ESs. Network routers are
the typical IS equipments. ES may implement the total ATN seven-layer stack as required,
while IS implements only the low-level layers.
The VDL subnetwork realizes the communication between avionics subnetwork and
ground subnetwork. It locates between the ATN ISs (avionics subnetwork router and
ground subnetwork router) and implements only the lower three layers of OSI reference
model.

ATN high level protocol

Application layer

Presentation layer

Session layer

Transport layer

Internetwork sublayer

SNAcP (ISO 8208)

LME DLS

MAC

D8PSK

VDL2 protocol

FIGURE 5.1
The VDL2 protocol stack in the ATN.
102 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Packet layer

DLS sublayer

VME entity DLE DLE DLE

1 2 n
LME LME LME
1 2 n

Queue
Data link layer

MAC sublayer

Physical layer

FIGURE 5.2
The VDL2 system architecture.

For the VDL2, the physical layer (at the bottom of system architecture) specifies a D8PSK
modulation scheme and operates at a bit rate of 31.5 kbps. All transmitted physical frames
contain a training sequence (Figure 5.2).
Above the physical layer is data link layer, which can be divided into MAC sublayer, DLS
sublayer, and the VDL management entity (VME). MAC is the interface between physical
layer and data link layer. It is governed by a p-persistent CSMA protocol, in which access
attempts are made only when the channel is sensed idle. During an access attempt, the sta-
tion will transmit with probability p or back off for TM1 seconds with a probability (1 − p).
The maximum number of access attempts is bounded, and after a maximum number of
failed access attempts, the MAC will transmit the packet as soon as the channel becomes
idle. This algorithm attempts to reduce the number of collisions while minimizing the
medium access delay.
The DLS employs a connection-oriented protocol (COP) known as aviation VHF link
control (AVLC). The AVLC protocol is a sliding window protocol with multiselective
reject functionality. It uses a dynamically calculated retransmission delay T1 based
on the channel utilization and the number of retransmissions. The value of T1 is ran-
domly set between a minimum and a maximum value, where the maximum is deter-
mined by the number of retransmissions. The DLS can support more than one data
link entity (DLE). All frames sent from DLEs are placed in an active DLS queue. The
queue processes frames for acknowledgment and redundancy. The queue guarantees
that redundant or duplicate packets are not contained in the queue. When a frame is
VDL2 Key Technology and Simulation 103

queued, its acknowledgment is checked, and all other frames to the same destination
are updated to reflect that acknowledgment. Incoming frames are also checked, and
any queued frames that the incoming frame acknowledges are removed. When the
access to the medium is successful, link control frames are transmitted in preference
to user data (UD). When the channel is accessed, more than one DLE frame may be
included in a single physical layer transmission.
VME provides link management functions, including link setup/release and handover.
It can support more than one link management entity (LME). Packet layer (subnet layer) is
mainly responsible for packet segment/reassembly, packet switching and error recovery,
subnet connection management, etc., functions.

5.2.2  Physical Layer

In baseband processing, the VDL2 applies the Nyquist first criterion with the raised cosine
roll-off filter (α = 0.6) to eliminate intersymbol interference. In band processing, the VDL2
applies D8PSK modulation scheme with the message rate of 31.5  kbps. In error correc-
tion, the VDL2 adopts the FEC mechanism and applies RS channel coding technology.
According to SARPs, scrambling and interleaving are also recommended to implement in
physical layer. The former can improve the synchronization performance of data stream;
the latter can enhance the capacity of anti-burst error by means of transforming channel.
In frequency allocation, the VDL2 assignable frequency covers the band
118–136.975  MHz. The separation between assignable frequencies (channel spacing) is
25 kHz, which can accommodate up to 760 channels. Frequency 136.975 MHz shall be
reserved as a worldwide common signaling channel (CSC) for the VDL Mode 2.

5.2.2.1 Architecture
As the stack bottom, the physical layer provides a physical link, which can realize trans-
parent transmission for communication. In the communication process between the
peers, the data link layer at the transmitter side, from high to low, transfers control sig-
naling and UD to the physical layer through service primitives. The physical layer will
implement coding/modulating onto the received message and allocate physical channel
according to upper layer request, then transmit the processed message to the physical
layer in the peer entity. At the receiver side, the physical layer transfers the received data
to data link layer through service primitives. In short, the physical layer provides service
for the data link layer in the local entity, and the peer layers realize physical layer proto-
col by the primitive-based services.

5.2.2.2 Functionality
The VDL2 physical layer can provide transceiver frequency control, bit exchange over
radio, notification, etc., functions, as follows:

1.
Transceiver frequency control: According to the link layer requests, selects transmit-
ting and receiving frequency.
2.
Notification: The GS notifies signal quality by signal quality parameter (SQP),
which is the basic for handover decision and link selection.
104 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

3.
Data transmitting: Encodes and modulates the data from the data link layer, trans-
mits the data via RF channel.
4.
Data receiving: Via RF channel, receives the data from the peer transmit-
ter, then decodes and demodulates the data. It is the inverse function of data
transmitting.
5.
Bit synchronization: The receiver extracts the synchronization information from the
transmitter and ensures the right time in data decision.

5.2.2.3  Physical Frame Structure

The VDL2 physical frame can be divided into six domains, the first five segments of which
construct training sequence (Table 5.3).
Each segment of training sequence has the meaning as follows:

1.
Transmitter ramp-up and power stabilization: The purpose of the segment, called the
ramp-up, is to provide for transmitter power stabilization and receiver AGC set-
tling, and it shall immediately precede the first symbol of the unique word. The
duration of the ramp-up shall be five symbol periods. The time reference point (t)
is the center of the first unique word symbol that occurs half a symbol period
after the end of the ramp-up. Conversely stated, the beginning of the ramp-up
starts at t = −5.5 symbol periods. The transmitted power shall be less than −40 dBc
prior to time t = −5.5 symbol periods. The ramp-up shall provide that at time
t = −3.0 symbol periods, the transmitted power is 90% of the manufacturer’s stated
output power or greater. Regardless of the method used to implement (or truncate)
the raised cosine filter, the output of the transmitter between times t = −3.0 and
t = −0.5 will appear as if “000” symbols.
2.
Synchronization code: The segment consists of the unique 48-bit word: 000 010 011
110 000 001 101 110 001 100 011 111 101 111 100 010, which implements the symbol
synchronization function.
3.
Reserved symbol: The segment consists of the single symbol representing 000 and is
reserved for future definition.
4.
Transmission length: To allow the receiver to determine the length of the final RS
block, the transmitter shall send a 17-bit word, from least significant bit (lsb) to
most significant bit (msb), indicating the total number of data bits that follow the
header FEC. From the field, we can know that the maximum length of physical
frame is 217–1 = 131,071 bits.
5.
Header FEC: To correct bit errors in the header, a (25, 20) block code shall be
computed over the reserved symbol and the transmission length segments. The

TABLE 5.3
VDL2 Physical Frame Structure
Transmitter Ramp-Up Synchronization Reserved Transmission Header
and Power Stabilization Code Symbol Length FEC Message
5 symbols 48 bits 3 bits 17 bits 5 bits Variable
VDL2 Key Technology and Simulation 105

encoder shall accept the header in the bit sequence that is being transmitted.
The parity bits to be transmitted shall be generated by the following equation,
where P is the parity symbol, R is the reserved symbol, TL is the transmission
length symbol, H is the parity matrix, and T represents the matrix transpose
function.

⎡ P1 P2 P3 P4 P5 ⎤ = ⎡ R1 R2 R3 TL1 … TL17 ⎤ PT (5.1)

⎣ ⎦ ⎣ ⎦

⎡ 00000000111111111111 ⎤
⎢ ⎥
⎢ 00111111000011111111 ⎥
P=⎢ 11000111001100001111 ⎥
⎢ 11011011010100110011 ⎥
⎢ ⎥
⎢⎣ 01101001111001010101 ⎥⎦

5.2.3  Data Link Layer

The VDL2 data link layer contains three parts: MAC sublayer, DLS sublayer, and the
VME.

5.2.3.1  Media Access Control

MAC sublayer applies P-CSMA and implements physical channel access and control.

5.2.3.1.1 Architecture
MAC sublayer shall accept the services from its lower layer, for example, it can acquire
channel idle/busy state from physical layer and use the bit in P-CSMA algorithm. It shall
provide the transparent acquisition for shared communication path and supply services to
the upper layer. As a whole, it can be regarded as the interface between the physical layer
and the data link layer.

5.2.3.1.2 Functionality
1. Data transmitting/receiving: MAC sublayer provides transparent transmission func-
tionality for upstream data and supplies P-CSMA-based access control for down-
stream data.
a. Transmitting process: MAC transfers AVLC frame from the DLS to physical
layer, which is a downward communication.
b. Receiving process: MAC transfers AVLC frame from physical layer to the DLS,
which is an upward communication.
2. Multiple access control: Based on P-CSMA algorithm, MAC sublayer can adminis-
trate and allocate the VHF physical media and can realize channel sharing and
multiplexing.
3. Channel congestion notification: When the channel is congested, MAC sublayer shall
send the channel congestion notification to VME, which can notify VME to initiate
a handoff.
106 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

5.2.3.2  Data Link Service

The DLS sublayer implements the AVLC protocol, mainly realizing frame sequencing,
error detection and recovery, and GS identification functions.

5.2.3.2.1 Architecture
The DLS sublayer is responsible for the organization and the transmission of packets on
the data link, implementing AVLC, which is derived from the HDLC.
The DLS sublayer shall build a DLE link for any point-to-point communication. A super-
vision relationship shall come into being between a local DLE and its associated remote
DLE, implementing the peer-to-peer communication control. At the same time, DLE is also
responsible for notifying packet-error information to VME.

5.2.3.2.2 Functionality
The main functions of the DLS sublayer are as follows:

1.
Frame sequence reception: In accordance with the serial number provisions, the DLS
sublayer can decide and receive AVLC frame.
2.
Flow control: Receiving side shall control the data transmission flow by sending RR
frame (RNR frame is not currently used) to the peer.
3.
Error detection: The DLS sublayer has the ability to detect and abandon the error
frame caused by transmission.
4.
GS address validation: AVLC frame has the unique source address and destination
address, which can realize the GS recognition.
5.
Retransmission: The DLS sublayer can retransmit an AVLC frame that is lost or
invalid in the destination.

5.2.3.2.3  AVLC Frame Structure

Inherited from the HDLC protocol, AVLC protocol has the similar frame structure. The
only difference between them exists in the address field. AVLC extends the address from
8 bits (HDLC frame) to 32 bits (both in source address and in destination address), which
can realize the transmitter authentication and enhance the system security (Figure 5.3).
In AVLC frame, the fields are defined as follows:

1.
Flag: The field shall identify the frame boundary and realize frame synchroniza-
tion. It is preset to 01111110.
2.
Address
a. Broadcast and multicast: When the address type is 1/1/1 and the other desti-
nation address bits are all 1, the VDL2 works in broadcast mode. When the
address type is not 1/1/1 (such as 1/0/0) and other destination address bits are
all 1, the VDL2 works in multicast mode.

Flag Address Control Information FCS Flag

8 bits 64 bits 8 bits varied lengths 16 bits 8 bits

FIGURE 5.3
AVLC frame structure.
VDL2 Key Technology and Simulation 107

b. Address type subfield: 001 represent 24-bit ICAO address space at AC. 100 and
101, respectively, represent the 24-bit ICAO-administered address space and
the 24-bit ICAO-delegated address space at GS. Among them, the ICAO-
administered address includes a country code prefix (using the same coun-
try code assignment defined in the ICAO Annex 10, Volume III, Chapter 9,
Appendix 1, table 1) and a suffix, which is assigned by appropriate author-
ity. The ICAO-delegated address is decided by the delegated organization.
111 represent the broadcast address space. The other formats are reserved for
future use.
c. A/G bit: The bit is used to specify the type of the message sender, where 0 rep-
resents aircraft (AC) and 1 represents GS.
d. C/R status bit: The status bit in the source address field shall be the command/
response (C/R) bit. The C/R bit shall be set to 0 to indicate a command frame
and set to 1 to indicate a response frame.
e. LSB (Least Significant Bit) bit: The bit is used to indicate the start and the end of
address field and is preset to 00000001.
3.
Control: The field is the same as the HDLC. According to the different frame types,
it can be divided into information (I) frame, supervisory (S) frame, unnumbered
(U) frame, etc., control field structure.
4.
FCS: Frame check sequence (FCS) field, generated by CRC polynomial, supports
the error detection function.
5.
Information: The field is consisted of UD, Initial Protocol Identifier (IPI), and
Extended IPI (ExIPI). UD shall bear the data from user with a variable length.
IPI and ExIPI can distinguish the message type (CLNP protocol message, ES–IS
protocol message, and IS–IS protocol message). According to the message type,
packet is delivered to the corresponding module (CLNP, IS–IS, and ES–IS)
(Table 5.4).

5.2.3.3  VDL Management Entity

As the highest level of data link layer, VME accepts services from the DLS sublayer and
physical layer. At the other side, VME provides services to network layer and can be
regarded as the interface between the data link layer and the upper layer.

5.2.3.3.1 Architecture
A VME shall have an LME link for each peer LME. Hence, that is, a ground VME shall
have an LME per aircraft, and an aircraft VME shall have an LME per ground system.
An LME shall establish a link between a local DLE and a remote DLE associated with its
peer LME. A ground LME shall determine if an aircraft station is associated with its peer
aircraft LME by comparing the aircraft address; two aircraft stations with identical aircraft
addresses are associated with the same LME. An aircraft LME shall determine if a GS is
associated with its peer ground LME by bit-wise logical ANDing the DLS address with
the GS mask provided by the peer ground LME; two GSs with the identical masked DLS
addresses are associated with the same LME.
Each aircraft and ground LME shall monitor all transmissions from its peer’s station(s)
to maintain a reliable link between some GSs and the aircraft while the aircraft is in the
coverage of a GS.
108 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 5.4
AVLC Frame Bit Structure
Bits
Description Bytes 8 7 6 5 4 3 2 1
Flag 0 0 1 1 1 1 1 1 0
Destination address 1 D22 D23 D24 D25 D26 D27 A/G 0
Address type
2 D15 D16 D17 D18 D19 D20 D21 0
3 D8 D9 D10 D11 D12 D13 D14 0
4 D1 D2 D3 D4 D5 D6 D7 0
Source address 5 S22 S23 S24 S25 S26 S27 C/R 0
Address type
6 S15 S16 S17 S18 S19 S20 S21 0
7 S8 S9 S10 S11 S12 S13 S14 0
8 S1 S2 S3 S4 S5 S6 S7 1
Link control 9 P/F
Information IPI 10 1 1 1 1 1 1 1 1
ExIPI 11 1 1 1 1 1 1 1 1
UD 12 … N − 2 User data
FCS N−1
N
Flag N + 1 0 1 1 1 1 1 1 0

5.2.3.3.2 Functionality
1. Frequency management: When the VDL2 is stated on a given service provider, the
VHF digital radio (VDR) is configured to filter uplinks which address matches
the provider prefix (10 first bits of the total 24 bits). VME shall scan up to 4 min
on CSC, on which GS regularly transmits broadcast frames (GSIF) to announce its
presence and position information. Every GSIF during the scan period initiates a
link establishment procedure with the uplink address.
2. GS management: VME stores all the information relative to the GS in coverage in a
peer entity connection (PEC) table and adds signal quality information provided
by the VDR with each uplink. VME can use PEC to make connection choice. And
if the signal quality of the current connected GS is too low, VME shall initiate a
handoff procedure to another GS, which can provide better signal quality.
3. Link-Change Notifications: The VME shall notify the intermediate-system system
management entity (IS-SME) of changes in link connectivity by supplying infor-
mation contained in the exchange identity (XID) frames received.

5.2.4  Subnetwork Layer

The subnetwork layer protocol is referred to formally as a subnetwork (across the VHF
air–ground subnetwork) access protocol and shall conform to ISO 8208 protocol. On the
AC/GS interface, the aircraft subnetwork entity shall act as data terminal equipment
(DTE) and the ground subnetwork entity shall act as data circuit-terminating equip-
ment (DCE).
VDL2 Key Technology and Simulation 109

5.2.4.1 Architecture
Data link layer shall strip off the layer 2 header and trailer from the received frame and
pass the remaining DLS UD up to subnetwork layer within a DLS primitive. This remain-
der is a subnetwork protocol data unit (SNPDU).

5.2.4.2 Functionality
1. Data control: Subnetwork layer is responsible for controlling data packet flow with
respect to duplicate, lost, or invalid SNPDU. Subnetwork layer can break SNPDU
into segments, which is called SNPDUs, for data control and error recovery.
2.
Virtual circuit: Subnetwork layer is basically responsible for internal routing (or
relaying) and provides data transferring across the subnetwork. It shall set up
and maintain the connection between air and ground routers, which is known as
virtual circuit, according to ISO 8208.

5.3  Key Technology

For the VDL2, some technologies will play key roles in bettering performance and have
been demanded or recommended by the VDL SARPs. Most of the key technologies are
widely used in present communication systems and are well known to the communication
engineer or researcher. But the key technologies often have their specialty when used in
an aviation air–ground communication system, where media is radio and bandwidth is
limited. In the section, several key technologies will be discussed and analyzed under the
VDL2 environment.

5.3.1  Physical Layer Key Technology

The principle of physical layer design is to select appropriate technologies according to
the real channel characteristics, which can, at system’s best, improve the reliability under
limited resources (such as power or bandwidth). The choice of technology is decided by
channel characteristics, that is, the strategy of physical layer technology is varied with
the difference of channels. Essentially, the fundamental purpose of applying the physical
layer technology is to enable the system to adapt to the physical channel.
Operating in VHF band, the VDL2 system has good propagation characteristics, which
can neglect multipath/selective fading effect and can be considered as having only electro-
magnetic wave attenuation with the distance. Therefore, there are no anti-fading measures
(such as spread spectrum, diversity reception) in the VDL2. In modulation, the VDL2 uses
the D8PSK scheme, which can improve the information transmission rate. In channel cod-
ing, the VDL2 uses the RS FEC coding, which can enhance the ability of anti-burst error
in mobile environment.

5.3.1.1  D8PSK Modulation

In the process of D8PSK implementation, the first step is to implement parallelization
(1/3), and next step is to map phase difference based on Gray code. The major character-
istic of Gray code is that the distance between adjacent code words is always 1, that is, if
110 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 5.5
Phase Mapping
Xk Yk Zk △φk
0 0 0 0π/4
0 0 1 1π/4
0 1 1 2π/4
0 1 0 3π/4
1 1 0 4π/4
1 1 1 5π/4
1 0 1 6π/4
1 0 0 7π/4

a bit error happens in Gray code, the original code word will be replaced by the adjacent
one. Obviously, a (Gray code) bit error will cause smaller error in sample receiving than
that of natural binary code (Table 5.5).
After the phase-difference mapping, differential encoding needs to be calculated from
the previous phase and the phase difference.

Φ k = Φ k−1 + ΔΦ k (5.2)

After the completion of phase encoding, the modulated signal can be obtained by IQ
orthogonal modulation.
If the correlation detection is applied for demodulation, the symbol error rate (Pe) under
big signal noise ratio (SNR) can be calculated by the following formula9:

⎛ 2Es π ⎞
Pe ≈ 2Q ⎜ sin ⎟ (5.3)
⎝ N0 2M ⎠

where
Es is the energy of a symbol
N0 is the noise power spectral density
M is the level of coding (for D8PSK, M = 8)

For phase modulation and Gray code, bit error rate (BER, Pb) can be calculated by the fol-
lowing formula:

Pe (5.4)
Pb ≈
log 2 M

From (5.3) and (5.4), we can get the BER as per the following equation:

1 ⎛ 2Es π ⎞ 2 ⎛ 2Es π ⎞
Pb ≈
log 2 M
2Q ⎜ sin ⎟ ≈ Q⎜ sin ⎟ (5.5)
⎝ N0 2M ⎠ 3 ⎝ N 0 8 2⎠

where Q function can be derived from error function or complementary error function.

1 ⎛ x ⎞ 1⎛ ⎛ x ⎞⎞
Q(x) = erfc ⎜ = 1 − erf ⎜ (5.6)
2 ⎝ 2 ⎟⎠ 2 ⎜⎝ ⎝ 2 ⎟⎠ ⎟⎠

VDL2 Key Technology and Simulation 111

Considering the following equation,

Pr E E
= b R = s Rs (5.7)
N0 N0 N0

we have

Rs Rs 1
Eb = Es = Es = Es (5.8)
R Rs log 2 M 3

where
Pr is the received power
Eb is the bit energy
R is the bit rate
Es is the symbol energy
Rs is the symbol rate

From (5.5) and (5.8), we can get the following equation, which can evaluate the BER perfor-
mance of D8PSK modulation:

1 ⎛ 6Eb π ⎞
Pb ≈
3
2Q ⎜ sin ⎟ (5.9)
⎝ N0 8 2⎠

5.3.1.2  RS Channel Encoding

In channel coding, the VDL2 applies RS code. As a q-nary code, RS code is the most impor-
tant subclass of BCH code. Each element value of RS code is decided by the Q element
symbol set:

{0, α , α ,…α } (5.10)
0 1 q−2

In the process of coding, q value is usually 2 to the power of a positive integer. As a

result, the entire nonzero elements (in the Q element set) belong to a mth degree primitive
polynomial-based extension field GF (2m), which represents the Galois field on the finite
set of integer 2m.10
In the VDL2 physical layer, channel coding applies RS (255,249) system code in which the
code length is fixed and m is 8. The primitive polynomial chosen is as follows:

p(x) = (x 8 + x 7 + x 2 + x + 1) (5.11)

The generation polynomial can be gotten as follows:

125

g(x) = ∏ (x − α ) (5.12)
i

i=120

where α is the primitive element (in the Galois field GF (28)), which is derived from (5.11).
112 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Due to q = 28 (256-nary), each symbol represents 1 octet and can transmit 8 bits. Therefore,
RS (255,249) will make block code for 249 octets (1992 bits) and will add 6 octets (48 bit)
redundancy bytes in coding.
The block number k is as follows:

Data length (bits) (5.13)

k=
1992 bits

The error-correcting capability can be calculated from the following formula:

⎡ ( dmin − 1) ⎤ ⎡ (n − k) ⎤ ⎡r⎤ ⎡ 48 ⎤
t = int ⎢ ⎥ = int ⎢ 2 ⎥ = int ⎢ 2 ⎥ = int ⎢ 2 ⎥ = 24 bit (5.14)
⎣ 2 ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦

The error-correcting ratio (ECCratio) is

Error-correctable message 3×8 1 (5.15)

ECCratio = = =
Total message 255 × 8 85

5.3.2  Data Link Layer Key Technology

5.3.2.1 P-CSMA
As a competitive access mode, P-CSMA inherits the essential features of CSMA and
increases the probability judgment algorithm.

5.3.2.1.1  Working Mechanism

P-CSMA keeps checking channel status and waits for the channel to be idle. When the
channel becomes idle and the transmitting number does not reach the maximum number
of access (M1), the system will transmit queued data with probability p or wait until the
next transmission timing with probability 1 – p. After reaching M1, the system will trans-
mit immediately under the idle state. If queued data are not transmitted after channel busy
timer (TM2) expiration, the VDL2 MAC sublayer will detect congestion and relay the mes-
sage to VME. If the channel is busy, the system continuously monitors until the channel
becomes idle. After becoming idle, the system will repeat the same as mentioned earlier.
According to whether the time slot is applied, P-CSMA can be classified into two modes11:

1.
Slot mode: Sending time is divided into many small intervals (slots), and media can
only be accessed in the front edge of a slot. In this mode, the nodes keep synchro-
nization via the master clock of the network.
2.
Nonslot mode: Sending time is not divided into slots, and media can be accessed at
any time. In this mode, the network synchronization is not needed (Figure 5.4).

5.3.2.1.2  MAC Sublayer Parameters

As the key technology of media access control, P-CSMA is closely relevant to the param-
eters regulated in the VDL2 MAC sublayer (Table 5.6).

1.
Parameter P: The parameter p (0 < p ≤ 1) is the probability under which the MAC
sublayer will transmit on any access attempt that will determine whether the
transmitter should be immediately enabled. An access attempt shall be made when
VDL2 Key Technology and Simulation 113

Packet queued

Channel Yes
busy?

No

Attempts Yes
P=1
>M1?

No

Increment No
Transmit ? Wait TM1
attempts

Yes

Packet sent

FIGURE 5.4
P-CSMA flowchart.

TABLE 5.6
MAC Sublayer Parameters
Description Unit Minimum Maximum Default Increment
M1 Maximum number of access Times 1 65,535 135 1
P Persistence No 1/256 1 13/256 1/256
TM1 Interaccess delay Millisecond 0.5 125 4.5 0.5
TM2 Channel busy Second 6 120 60 1

Timer TM1 expires and the channel is idle, or when a transmission request arrives
from the DLS while the channel is idle, or if the channel is determined to become
idle while a message is queued for transmission. The result of an access attempt
is decided by the random number produced by P-CSMA: if the random number
is less than P, the access attempt is successful; otherwise, the access attempt is
unsuccessful. If the access attempt is successful, the transmission shall immedi-
ately begin.
2. Counter M1: The M1 counter shall evaluate the maximum number of attempts (M1)
that a MAC sublayer will make for any transmission request. This counter shall
be started upon system initialization, Timer TM2 expiring, or a successful access
attempt. The counter shall add 1 after every unsuccessful access attempt. When
the counter expires, authorization to transmit shall be granted as soon as the chan-
nel is idle.
3. Timer TM1 (interaccess delay timer): The TM1 timer shall be set to the time (TM1)
that a MAC sublayer will wait between consecutive access attempts. This timer
114 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

shall be started if it is not already running, and the channel is idle after an access
attempt is unsuccessful. The timer shall be canceled if the channel becomes busy.
When the timer expires, another access attempt shall be made.
4. Timer TM2 (channel busy timer): The TM2 timer shall be set to the maximum time
(TM2) that a MAC sublayer will wait after receiving a request to transmit. This
timer shall be started, if it is not already running, when the MAC sublayer receives
a request for transmission. The timer shall be canceled upon a successful access
attempt. When the timer expires, the VME shall be informed that the channel is
congested.

5.3.2.2  From HDLC to AVLC

AVLC sublayer conforms to HDLC as specified by ISO 13239.12 However, given that HDLC
was designed to primarily support stationary network terminals where bandwidth for the
most part is not scarce, AVLC has been optimized to consider the fact that the VDL net-
work terminals are in a mobile environment with limited bandwidth available and to add
the adaptive channel estimation algorithm in packet retransmission.

5.3.2.2.1  HDLC Mechanism

In 1974, IBM launched a famous system structure, SNA, in which data link layer protocols
use a bit-oriented Synchronous Data Link Control (SDLC). Later, ISO modified SDLC to
HDLC as international standard.
Applying serial number and confirmation, HDLC can provide reliable transmission
with error free characters, no lost or repeated packets. The HDLC frame is constructed
from address field, control field, information field, FCS, and two flags (Figure 5.5).
According to the difference of functionality, the HDLC frame can be divided into three
categories, as shown in Table 5.7.

1.
Information frame (I frame): I format is used to perform an information transfer.
2.
Supervisory frame (S frame): S format is used to perform data link supervisory
control functions such as acknowledging I frames, requesting retransmission of
I frames, and requesting a temporary suspension of transmission of I frames.

8 bits 8 bits 8 bits Varible 16 bits 8 bits

Frame checking
Flag Address Control Information Flag
sequence
(F) (A) (C) (I) (F)
(FCS)

FIGURE 5.5
HDLC frame.

TABLE 5.7
Control Field Structure
1 2 3 4 5 6 7 8
Information frame (I) 0 N(S) P/F N(R)
Supervisory frame (S) 1 0 S P/F N(R)
Unnumbered frame (U) 1 1 M P/F M
VDL2 Key Technology and Simulation 115

3.
Unnumbered frame (U frame): U format is used to provide additional data link con-
trol functions and unnumbered information transfer. This format shall contain
no sequence numbers, but shall include a P/F bit. Five modifier bit positions are
available, thus allowing definition of up to 32 additional command functions and
32 additional response functions.

The functions of N(S), N(R), and P/F are independent:

1.
N(R) symbols transmitting receive sequence number (bit 6 = low-order bit), which
can indicate that the station transmitting the N(R) has correctly received all I
frames numbered up to [N(R) −1] (inclusive).
2.
N(S) symbols sending sequence number (bit 2 = low-order bit).
3.
P/F is the poll or final bit: The poll (P) bit set to “1” shall be used by the primary/
combined station to solicit (poll) a response or sequence of responses from the
secondary station(s)/combined station. A response frame with the F bit set to “1”
shall be used by the secondary/combined station to acknowledge the receipt of a
command frame with the P bit set to “1.”

To reduce the times of acceptance acknowledgment and to improve the transmission effi-
ciency, sliding window and piggybacking are applied in the HDLC procedure:

1. Within the window length (default value is 7), communication peers can continu-
ously send I frame, which is not acknowledged by the peer.
2. Acknowledgment can be piggybacked by I frame from the peer, and the frames
from 0 to [N(R) −1] (mode 8) can be confirmed by N(R) bit at once.

To avoid the transmitter waiting acknowledgment for an undesirable lag, HDLC intro-
duces the time-out retransmission: whenever the transmitter sends an I frame, it will start
the retransmission timer and will keep timing until receiving confirmation (including pig-
gybacking) from the peer. When a frame is correctly received, the receiver will initiate its
timing process. The timer will be canceled when the receiver has sent at least a frame that
piggybacks the acknowledgment information. At the expiration of the receiver timer, an S
frame will be sent to inform the transmitter that the transmitted frames has been received
correctly.

5.3.2.2.2  AVLC Improvement on HDLC

Originating from HDLC, AVLC has the inheritance with HDLC and can be regarded
as a subset of HDLC. It applies the adaptive channel estimation algorithm in the HDLC
retransmission, which is the most different point from HDLC and the most prominent
improvement on HDLC. The algorithm is proposed to adapt the characters of terminal
mobility and frequency scarcity existing in the air–ground data link control.
HDLC uses a constant timing retransmission in which there exists its shortcoming:
on one hand, when the channel utilization rate is comparably high, retransmission
frame after a constant delay before retransmission (T1) is easy to collide in P-CSMA
and will further aggravate the channel burden. On the other hand, when the channel
utilization rate is comparably low, retransmission frame after a constant T1 will lead to
unnecessary delay. The shortcomings become obvious and unbearable in a frequency-
scarce system.
116 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 5.8
T1 Parameters
Lower Upper
Symbol Parameter Name Bound Bound Default Increment
T1min Delay before Minimum 0s 20 s 1.0 s 1 ms
T1max retransmission Maximum 1s 20 s 15 s 1 ms
T1mult Multiplier 1 2.5 1.45 0.01
T1exp Multiplier 1 2.5 1.7 0.01

To overcome the shortcomings, AVLC applies the adaptive channel estimation algorithm
for T1 timer. It is based on the idea that radio channel has a continuous character and the
future channel utilization rate can be predicted from that of the current. In the algorithm,
the predicted value is applied to adjust T1 by increasing it with a smaller channel utiliza-
tion rate and decreasing it with a bigger channel utilization rate.
The T1 adaptive channel estimation algorithm is as follows:

( )
T1 = T1 min + 2 × TD99 + min u(x),T1 max (5.16)

where

TM1 × M1
TD99 = (5.17)
1−µ

In (5.17), TD99 is the running estimate for the 99th percentile transmission delay (between
the time at which the frame is sent to the MAC sublayer and the time at which its transmis-
sion is completed). Parameter μ is a measurement of channel utilization with a range of
value from 0 to 1, where 1 corresponding to a channel that is 100% occupied. TM1 and M1
are parameters of P-CSMA.
The parameter u(x) is a uniform random number generated between 0 and x that is cal-
culated from the following formula:
retrans
x = T1mult × TD99 × T1exp (5.18)

In (5.18), the superscript retrans is the largest retransmission count of all the outstanding
frames.
T1mult, T1exp, T1min, and T1max are configurable parameters, which are listed in Table 5.8.
From the earlier algorithm, we can see that with the channel utilization (μ) increasing,
transmission delay (TD99) and x will increase. With the increase of x, u(x) will increase in
the probability of uniform distribution. Moreover, retransmission (T1) will increase with
the probability growing up. As a result, T1 and μ will change in the same direction, that is,
retransmitted frames will be sent slower when channel utilization rate is higher; retrans-
mitted frames will be sent faster when channel utilization rate is lower.

5.3.2.3 Handoff
An aircraft flies along its flight path, and it will move into or out of range of various VDL
GSs, which will lead to the change of signal quality. To provide continuous data link
VDL2 Key Technology and Simulation 117

communications, the aircraft radio must switch its linking from the current GS to a new
GS, which can supply a better signal. The process is called handoff.
The VDL2 supports four variations of the handoff, including aircraft-initiated handoff,
aircraft-requested and ground-initiated handoff, ground-initiated handoff, and ground-
requested and aircraft-initiated handoff. Aircraft-initiated handoff should be the usual
mechanism for a light to moderately loaded subnetwork. Ground-initiated handoff and
ground-requested and aircraft-initiated handoff occur under heavy loading conditions to
perform load balancing between GSs. A ground-initiated handoff must be to another GS
on the same frequency. Therefore, the ground-requested and aircraft-initiated handoff is
included to provide for frequency changes. An aircraft-requested and ground-initiated
handoff is provided to allow an aircraft to decide when a handoff should occur but allows
the ground network to decide which is the best GS for load balancing.13

5.3.2.3.1  Aircraft-Initiated Handoff

Once the aircraft LME has established a link to a GS, it shall monitor the VHF signal qual-
ity on the link and the transmissions of the other GSs. The aircraft LME shall establish a
link to a new GS due to the following reasons:

1.
Signal quality: The VHF signal quality on the current link is poor, and the signal
quality of another GS is significantly better.
2.
Link failure
a. The counter N2 is exceeded on any frame sent to the current GS.
b. The timer TG2 expires for the current link.
c. The timer TM2 expires.

In the way of aircraft-initiated handoff, an aircraft decides when to start initiating hand-
off and sends an XID_CMD_HO message to the newly chosen GS in order to request the
handoff. If the request is accepted, the GS will respond with an XID_RSP_HO message;
otherwise, the GS will reject the proposed handoff with an XID_RSP_LCR message. When
a handoff has been accepted, the aircraft terminates the old data link (1) by transmitting a
DISC frame, perhaps after waiting for a short time to ensure that all data have been trans-
ferred over the old data link. This procedure is performed when handoff occurs between
service providers or (2) by both avionics and the old GS starting a timer (TG5) and auto-
matically clearing the calls at timer expiry. This procedure is used when handoff occurs
between GSs belonging to the same service provider and implies that some sort of mes-
sage is sent between new and old GSs (Figure 5.6).

5.3.2.3.2  Ground-Initiated Handoff

In the way of ground-initiated handoff, an aircraft decides when to start initiating handoff
and sends an XID_CMD_HO message to an aircraft in order to request the handoff. If the
request is accepted, the aircraft will respond with an XID_RSP_HO message. The handoff
decision and TG5 timing of this mode are similar to aircraft-initiated handoffs and will
not be described here.

5.3.2.3.3  Aircraft-Requested and Ground-Initiated Handoff

In the way of aircraft-requested and ground-initiated handoff, an aircraft first sends an
XID_CMD_HO message to a GS and does not require immediate response. The GS will
hand over the message, via ground–ground communication, to the next GS, which will
118 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

1. XID_CMD_HO

2. XID_RSP_HO 3. Start TG5

3. Start TG5 LME 4. Disconnect

timer old GS at TG5
timer expiry
4. Disconnect old GS
at TG5 timer expiry
AC New GS Old GS

FIGURE 5.6
The process of aircraft-initiated handoff.

decide when to start initiating handoff and will send an XID_CMD_HO (P = 1) message
to the aircraft. If the request is accepted, the aircraft will respond with an XID_RSP_HO
(F = 1) message. The handoff decision and TG5 timing of this mode are similar to aircraft-
initiated handoffs and will not be described here.

5.3.2.3.4  Ground-Requested and Aircraft-Initiated Handoff

In the way of ground-requested and aircraft-initiated handoff, a GS first sends an XID_
CMD_HO message to an aircraft and does not require immediate response. The aircraft
will decide which GS to select and when to start initiating handoff. Then, it will send an
XID_CMD_HO (P = 1) message to the selected GS. If the request is accepted, the GS will
respond with an XID_RSP_HO (F = 1) message. The handoff decision and TG5 timing of
this mode are similar to aircraft-initiated handoffs and will not be described here.

5.3.3  Subnetwork Link and Connection

The VDL2 subnetwork layer is an implementation of ITU-T X.25 as specified in ISO 8208
and can supply a variety of facilities to establish, terminate, and manage connections.

5.3.3.1  Explicit Subnetwork Connection

The connection of an explicit subnetwork is based on the prerequisite of link establishment,
which is implemented by LME. If no link remains, an aircraft transmitting/receiving a DM
frame shall initiate link establishment with a GS. The aircraft LME shall choose a GS, with
which it wishes to establish a link based on the signal quality of all received uplink frames
and on information in any received GSIFs, and then shall attempt to establish a link with
the chosen GS by sending an XID_CMD_LE (P = 1) message. If the ground LME receives
the message, it shall confirm the link establishment by sending an XID_RSP_LE (F = 1) mes-
sage and complete the link procedure.
During link establishment, a ground DCE shall indicate its available routers in the ATN
outer NETs parameter, and the aircraft LME shall then attempt to maintain all subnetwork
connections.

1.
Explicit subnetwork connection establishment: Immediately after link establishment,
the aircraft DTE shall attempt to establish a subnetwork connection to at least
VDL2 Key Technology and Simulation 119

one ground DTE. The aircraft DTE shall request a single subnetwork connection
per ground DTE by the transmission of a CALL REQUEST packet specifying the
ground DTE address. On receipt of the CALL REQUEST, the ground DCE shall
attempt to establish a subnetwork connection to the aircraft DTE by responding
with a CALL CONFIRMATION packet; otherwise, the ground DCE shall send
a CLEAR REQUEST packet including the clearing cause and diagnostic code
of the failure. If Ground Network X.121 DTE addressing is implemented, the
ground DCE shall use the Called Line Address Modification Notification facility
to inform the aircraft DTE of the ground DTE’s X.121 address. Else, if the default
ground DTE addressing is implemented, the ground DCE shall use the Called
Address Extension facility to inform the aircraft of the ground DTE’s Ground
VDL-Specific DTE Addressing (VSDA) address that was delivered in the CALL
REQUEST.
2. Explicit subnetwork connection maintenance: To explicitly request subnetwork connec-
tion maintenance to a ground DTE, an aircraft DTE shall send a CALL REQUEST
packet to the ground DTE with the fast select facility set containing a VDL mobile
SNDCF Call UD Field indicating a request to maintain SNDCF context. If the
ground DTE can accept the call, it shall respond with a CALL CONFIRMATION
packet with the fast select facility set containing a VDL mobile SNDCF Call UD
field indicating whether the SNDCF context was maintained. If the ground DTE or
a DCE is unable to accept the call, it shall send a CLEAR REQUEST packet to the
aircraft DTE including the clearing cause and diagnostic code of failure. If Ground
Network X.121 DTE addressing is implemented, then the ground DTE shall use
the Called Line Address Modification Notification facility to inform the aircraft
DTE of the ground DTE’s X.121 address. Else if the default ground DTE address-
ing is implemented, then the ground DCE shall use the Called Address Extension
facility to inform the aircraft of the ground’s VSDA address that was delivered in
the CALL REQUEST.

5.3.3.2  Expedited Subnetwork Connection

To perform an expedited subnetwork connection establishment or maintenance, the
initiating LME shall include in the XID_CMD the Expedited Network Connection
parameter for each subnetwork connection that needs to be established or maintained.
If the responding LME receives an XID_CMD with one or more Expedited Network
Connection parameters, it shall confirm subnetwork connection establishment or main-
tenance by sending an XID_RSP containing the parameters. This function shall be
applicable only for the link establishment, air-initiated handoff, and ground-initiated
handoff processes.

1.
Expedited subnetwork connection establishment: The aircraft DTE shall reissue CALL
REQUESTs for those logical channels for which responses (i.e., either a CALL
CONFIRMATION or a CLEAR REQUEST) were not included in the XID_RSP_LE.
If Ground Network X.121 DTE addressing is implemented, then the ground DCE
shall use the Called Line Address Modification Notification facility to inform
the aircraft DTE of the ground DTE’s X.121 address. Else, if the default ground
DTE addressing is implemented, the ground DCE shall use the Called Address
Extension facility to inform the aircraft of the ground’s VSDA address that was
delivered in the CALL REQUEST.
120 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

2.
Expedited subnetwork connection maintenance: The initiating DTE shall reissue
CALL REQUESTs for those logical channels for which responses (i.e., a CALL
CONFIRMATION or a CLEAR REQUEST) were not included in the XID_RSP_HO.
A ground DTE shall include its Calling Address in the appropriate field. If Ground
Network DTE addressing is implemented, then the ground DTE shall use the
Called Line Address Modification Notification facility to inform the aircraft DTE
of the ground DTE’s X.121 address. Else, if the default ground DTE addressing is
implemented, the ground DCE shall use the Called Address Extension facility to
inform the aircraft of the ground’s VSDA address that was delivered in the CALL
REQUEST.

5.4  Modeling and Simulation

In this section, a VDL2 model, which includes three-layer protocol stack and aircraft/GS
nodes, is established on OPNET platform. Based on the established model, simulations are
carried out to analyze and optimize the performance of the data link.

5.4.1  VDL2 Modeling

Aircraft node and GS node are two kinds of nodes existing in the VDL air–ground data
link. They are the peers in the half-duplex communication, in which only one can transmit
message to the other at any time. They also have a similar protocol structure. As shown
in Figure 5.7, the node model can be divided into three layers from the top to the bottom
corresponding to the VDL2 protocol stack: (1) packet layer (subnetwork layer) constructed

FIGURE 5.7
(See color insert.) The GS node model.
VDL2 Key Technology and Simulation 121

from PK_gen module, (2) data link layer constructed from DLE/VME and MAC modules,
and (3) physical layer constructed from PK_gen module, which can implement the switch
between transmitter and receiver.
In modeling, the VDL2 channel can be divided into message channel, which is respon-
sible for point-to-point communication, and CSC channel, which is responsible for broad-
casting in the band within 136.975–137 MHz. A transmitter/receiver module will occupy
only a message or CSC channel and will be set up according to the channel type. Also, a
peer of transmitter and receiver work in the same frequency within a 25 kHz band. On the
bottom of aircraft node model is the omnidirectional antenna, which can communicate
with transmitter and receiver in the way of packet stream.
The difference between AC and GS nodes mainly lies in that aircraft can implement
handoff between multiple GSs. Accordingly, the aircraft side adopts a multiple message-
transceiver scheme, which is different from the GS side with the single message transceiver.

5.4.1.1  Physical Link Layer Modeling

The physical link layer is modeled in PK_gen module and transmitter/receiver module
and radio pipeline, implementing modulation, channel encoding, and physical frame
framing and switching.

1.
D8PSK modulation: In OPNET, modulation scheme is realized in a modulation
curve, which can reflect the relationship of BER and Eb/N0. Therefore, we realized
D8PSK modulation by Equation 5.9 using the modulation curve editor.
2.
RS encoding: The RS encoding error-correcting ratio can be calculated in Equation
5.15 and is set as the receiving SNR threshold in the error correction pipeline stage.
The key code is as follows:

/* Test if bit errors exceed threshold. */

if (pklen = = 0)
accept = OPC_FALSE;
else
accept = ((((double) num_errs)/pklen) < = ecc_thresh) ? OPC_TRUE:
OPC_FALSE;
/* Set flag indicating accept/reject in transmission data block. */
op_td_set_int (pkptr, OPC_TDA_RA_PK_ACCEPT, accept);

3.
Physical frame framing: In the VDL2, the data link layer frame is called the AVLC
frame, and the physical layer frame is called the physical frame. From top to bottom,
the AVLC frame will be encoding, header-adding, and packaging, and finally be con-
verted into the physical frame. The physical frame is the basic data unit to transmit in
the physical media and carries the node location, power, conflict, etc., information.
The frame can be formed in the frame editor. The physical frame model is as
given in Figure 5.8.
4.
Physical packet switching: For a sending packet, the task of physical frame switch-
ing is to realize the mapping from sending address to transmitter. For a received
packet, the task is to distinguish whether the destination address (of the received
packet) is consistent with the node address: if it is inconsistent, the received packet
will be destroyed; if it is consistent, the mapping from the address to data flow will
be implemented.
122 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

FIGURE 5.8
Physical frame model.

The mapping relationship is achieved through the packet stream ID. The key code is as
follows:

/* Packet communication for frame being sent in aircraft side. */

if(op_intrpt_strm() = =4)
op_pk_send(radio_pk,0);
if(op_intrpt_strm() = =5)
op_pk_send(radio_pk,2);
if(op_intrpt_strm() = =6)
op_pk_send(radio_pk,4);

To facilitate switching modeling, one GS is allocated only one 25 kHz channel, and the
channels are differed between different GSs. By this way, the transceivers of GS and air-
craft compose a one-to-one mapping, which makes the switching simplified.

5.4.1.2  Data Link Layer Modeling

5.4.1.2.1  MAC Module
The key of MAC process modeling is to realize P-CSMA protocol, including TM1 timer and
P probability decision.14

1.
GS MAC modeling: As shown in Figure 5.9, state INIT fulfills the tasks of variable
initialization, object parameters acquirement, statistic registration, and TM1 timer
initialization. DATA_CHAN reads Channel busy/idle statistics from physical
layer and updates message to the media access. State TM1_intr_des implements
TM1 timer.
State STORE realizes data stream shunting function with the application of
queue. The received data stream is transmitted to the upper layer in a transparent
way. According to the type of information, data packets are delivered into dif-
ferent queues waiting for processing. DATA_ACCE implements the algorithm of
P-CSMA and realizes the probability decision as follows:

/* P probability decision */
p = op_dist_uniform(1.);
if((channel_status = =IDLE)&&(p< = csma_p||counter_m1> = csma_nb_max))
{Message is dealt and sent.}
VDL2 Key Technology and Simulation 123

FIGURE 5.9
(See color insert.) The GS MAC process.

The State type of CSC is in symmetry with that of DATA and realizes processing
and transmitting of signaling in CSC channel.
2.
AC MAC modeling: The modeling method of AC MAC is similar to that of GS MAC.
The main difference exists in that not only one GS can be connected and switched
to an AC. As a result, there are multiple channels in an AC model.

In realizing the conversion of multiple frequencies, AC modeling applies the mechanism

of dynamic process. According to the current connection, the parent process dynami-
cally creates and destroys the child process, which is corresponding to the channel with
a fixed frequency. In MAC model of AC side, the parent process is responsible for chan-
nel assignment, process invoking, and CSC signaling processing. The child process is
responsible for implementation of the functions of a fixed channel such as UD sending
and probability decision.

5.4.1.2.2  DLS Module

The key of the DLS process modeling is to realize AVLC protocol. In the DLS model, the
GS side and aircraft side are all constructed with multiple DLE, single LLC, and sending
queue. A DLE corresponds to a logical link between GS and AC, which implements flow
control and deals with I frame and S frame. LLC is responsible for the control of logical
link and deals with U frame.
124 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

In the DLS modeling, we applied dynamic process, queue management, etc., OPNET
mechanisms.15

1.
Dynamic process: As the father process, the DLS process is first set up. It will be
responsible for dynamically building DLE that is taken as the child process. When
a logical link is established, the DLS will birth a DLE; when the link is released, the
child process will be destroyed.

//create and invoke child DLE_process.

if(op_pro_valid(DLE_proc) = =OPC_FALSE)
{
DLE_proc = op_pro_create(“VDL_aircraft_DLE_child_upd”,​
OPC_NIL);
}
op_pro_invoke(DLE_proc,OPC_NIL);

2.
Queue management: In the DLE child process, the I frame waiting to send is inserted
into a FIFO transmission queue. By sliding window, the first frame in the queue
will be sent to the MAC. The key code is as follows.

/* Remove the head from the subqueue and set the NS/NR bits. At the
same time, increase the current window code. */
s_pkptr = op_subq_pk_remove(queue_id,OPC_QPOS_HEAD);
op_pk_nfd_set(s_pkptr,“NS”,local_ns);
op_pk_nfd_set(s_pkptr,“NR”,local_nr);
m++;

5.4.1.2.3  VME Module

In the VME module, we realized handoff functionality based on dynamic signaling, deci-
sion timing, mobile node, multifrequency access, etc., methods.

1.
Dynamic signaling: From Section 5.3.2.3, we know that handoff will be implemented
via signaling interaction between AC and GS. We realized the process by finite-
state machine (FSM), using a dynamic process that can simplify the FSM structure
and improve the simulation efficiency.

FIGURE 5.10
(See color insert.) The AC handoff child process.
VDL2 Key Technology and Simulation 125

TABLE 5.9
Frequency Configuration
GS Lower Limit (MHz) Upper Limit (MHz) Band (kHz)
node_0 136.000 136.025 25
node_1 136.325 136.350 25
node_2 136.650 136.675 25

Dynamic signaling interaction asks for a strict logical relationship: only after hav-
ing received the signaling from peer, the local VME can send its own signaling
out. In order to guarantee the correct logic, we used a method of stream interrupt
combined with unforced state. As in Figure 5.10, the AC child process first sends
the XID_CND_HO command and then returns to unforced state. It will wait for
response from GS and enter into sleep. If the GS accepts the handoff request and
sends the command response, the response will be via packet stream sent to the
parent process and trigger the stream interruption. The parent process will wake
up the child process. The child process will make further processing and then
transfer to the next signaling state.
2.
Decision timing: The decision of handoff requires a TG5 timer, which is realized via
self interrupt in VME module. The key code is as follows:

//define transition from the unforced state Idle to the forced state
HO_destroy.
#define HO_destr op_intrpt_type() = =OPC_INTRPT_SELF&&op_intrpt_
code() = =TG5_int
//init_TG5 is the default value of TG5 in VDL2 protocol.
#define init_TG5 20
op_intrpt_schedule_self(op_sim_time()+init_TG5, TG5_int);

3.
Mobile node: Different from the fixed node GS, the AC node is mobile with high
speed. The mobility is closely related to the handoff process and is realized via
trajectory with segment-based variable interval in modeling.
4.
Multifrequency access: In the handoff process, AC will possibly switch between GSs
with different frequencies. To simplify the simulation, one GS is allocated one
single frequency that is far away from the others as possible to weaken the effect
of channel cross talk in simulation. The frequency configuration is implemented
in the GS antenna module as in Table 5.9.

5.4.1.3  Packet Layer Modeling

Since the VDL2 simulation focuses on the performance of the air–ground data link, the
ATN higher level and interface are not involved. Therefore, we take only the packet layer
as a data source, that is, a traffic model, ignoring the ATN concerned function. The traffic
model can provide data flow to the lower layer and is the basic means of verification and
simulation of the network. The design of traffic model should be scientific to ensure a rea-
sonable and reliable simulation.
In the VDL2, the traffic has burst characteristics in packet strength and random charac-
teristic in packet length. Accordingly, we abstracted the traffic model with two stochastic
126 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

processes, respectively: (1) packet strength modeled with Poisson distribution and (2)
packet length modeled with uniform distribution.

5.4.1.3.1  Stochastic Processes

5.4.1.3.1.1  Packet Strength
Definition 5.1: Let XT be a random process. For any n(n ≥ 2) and ti(t0 < t1⋯ < tn), if the incre-
mental X(t1) − X(t0), X(t2) − X(t1), … ,X(tn) − X(tn−1) are independent of each other, then XT is
called an independent increment process.

Definition 5.2: Let XT be a random process,  XT = {N(t), t ∈ T, T ∈ [0, ∞)}. If N(t) is a nonnega-
tive integer and N(s) < N(t) when s < t, then XT is called a counting process.

Definition 5.3: If the counting process XT = {N(t),  t ≥ 0} satisfies the following three
conditions:

1. N(0) = 0
2. XT is an independent increment process
3. For any nonnegative s and t, there is the equation

(λt)k
P { N(s + t) − N(s) = k } = e −λt k = 0,1, 2,…; λ > 0 (5.19)
k!

then, XT is called a Poisson process with Poisson strength λ.16

With the burst characteristics, packet birth can meet the following conditions:

1. In the nonoverlapping period, the arrivals of packets are independent.

2. The probability of packet arrival is only proportional to Δt and has nothing to do
with the starting point.
3. At any small enough period, the number of arrival packets is not more than one;
otherwise, it will be considered as impossible event.

With these conditions, we can draw a conclusion that the birth of burst packets meets the
Poisson process.

Proof:

1. From the three conditions of burst packets, it is easy to know that the birth of burst
packets is both an independent increment process and a counting process.
2. For any Δt, we can know that the arrival probability of a packet is Δt from con-
dition (2) and the nonarrival probability of a packet is 1 − Δt from condition
(3). Assuming that k packets are birthed during period t, we can divide t into n
little Δt, which is small enough and contains not more than one packet.

Let

Pk (s,t) = P{N(s + t) − N(s) = k}

VDL2 Key Technology and Simulation 127

we have

n!
Pk (s,t) = lim Cnk ( λΔt ) (1 − λΔt ) ( λΔt ) (1 − λΔt )
k n−k k n−k
= lim
n→∞ n→∞ k!(n − k)!
k n−k k n−k
n! ⎛ t⎞ ⎛ t⎞ (λt)k n! ⎛ 1 ⎞ ⎛ t⎞
= lim ⎜ λ ⎟ ⎜ 1 − λ n ⎟⎠ = lim ⎜ ⎟ ⎜ 1− λ ⎟
n→∞ k!(n − k)! ⎝ n ⎠ ⎝ k! n→∞ (n − k)! ⎝ n ⎠ ⎝ n⎠

(λt)k n! (n − λt)n−k
Pk (s,t) = lim n (5.20)
k! n→∞ n (n − k)!

Due to n → ∞, n − k → ∞, we have
n
n ⎛ λt ⎞
n! (n − λt)n−k ⎛ λt ⎞ ln ⎜ 1− ⎟
lim n = lim ⎜ 1 − ⎟ = lim e ⎝ n ⎠ = e −λt (5.21)
n→∞ n (n − k)! n→∞ ⎝ n⎠ n→∞

From Equations 5.20 and 5.21,

(λt)k −λt
Pk (s,t) = e (5.22)
k!

5.4.1.3.1.2  Packet Length  In the VDL2, the AVLC packet length l is random and can be
presumed to meet uniform distribution.

⎧ 1
⎪ a≤l<b
f (l) = ⎨ b − a (5.23)
⎪ 0 l < a or l ≥ b
⎩

5.4.1.3.2  Traffic Model

From random distribution theory, we know that if the packet arrival interval obeys expo-
nential distribution with the mean 1/λ, the packet sending process is a Poisson process
with strength λ and vice versa. According to the earlier analyses, we made the traffic
model flowchart as follows (Figure 5.11):

Random Packet Send to

packet length birth AVLC

Packets
interval

FIGURE 5.11
Traffic model flowchart.
128 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

In modeling, random functions are called

/* the function to create uniform distribution */

op_dist_uniform(msg_max_length-msg_min_length)
/* the function to create exponential distribution */
op_dist_exponential(iat)

Also, self interrupt is applied to trigger the next packet birth and is set to the sum of
current moment and exponential distribution-based interval:
next_pk_time = op_sim_time()+op_dist_exponential(iat);
op_intrpt_schedule_self(next_pk_time,NEW_PK_TIME);

5.4.2  Handoff Simulation

Based on the VDL2 model, simulations are carried out on throughput and delay. The
results are similar to the current research. For example, given that payload is defined as
the ratio of throughput and system transition rate (31,500 bits/s), we made the throughput
versus payload simulation on the model. The result (Figure 5.12) is in consistency with
Bretmersky and Kerczewski’s experimental result.17
In the section, we will focus on the VDL2 handoff and propose a handoff decision
method based on interference plus noise ratio (SINR). Simulations will be carried out to
comprehend the performance of the method.18

5.4.2.1  SINR-Based Handoff

From Section 5.3.2.3, we know that there exist four handoff mechanisms in the VDL2:
aircraft-initiated handoff is for the moderately loaded subnetwork; ground-initiated
and ground-requested aircraft-initiated handoffs occur under heavy loading conditions;

0.55

0.50

0.45

0.40

0.35
Throughput

0.30
0.25

0.20
0.15

0.10

0.05

0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Payload

FIGURE 5.12
Throughput versus payload.
VDL2 Key Technology and Simulation 129

aircraft-requested ground-initiated handoff provides ground decision for load balancing.

However, how to choose an appropriate mode is a complex and difficult task under a spec-
ified condition. It requires a rational and practical switching algorithm that has switching
overhead itself. In fact, aircraft-initiated handoff is the usual mechanism in the current
VDL2 handoff application.
We try to find a way to optimize aircraft-initiated handoff with the goal of (1) keeping
reliable and stable signal strength, (2) balancing the subnetwork load between GSs in a
continuous way, (3) considering the handoff overhead, and (4) using a simple algorithm.

5.4.2.1.1  SINR Criteria

As we know, signal attenuation is inversely proportional to the n square of propagation
distance, in which n depends on the transmission medium. Due to the distance effect,
the signal strengths will change inversely when an aircraft flies from one GS to another
GS. After the changing arrives at a certain threshold, a handoff has to be initiated from
the original GS to a new GS to keep a good signal quality. Here, signal strengths can be
described with SNR, due to the fact that the noise in the VDL2 channel can be regarded as
Gauss white noise, and the frequency band is fixed.
However, there exists another effect in the VDL2 network: the distribution of high-speed
nodes will change in a constant way, which will lead to the nodes density uneven. For the
competition-based P-CSMA, high density will cause an increase in media access delay.
Besides, uneven density will cause some channels to be too busy as well as some channels
to be too idle and result in an overall system capacity decline.
However, in P-CSMA, multiple nodes will access the channel in a preemptive way. The
media access process can be regarded as co-frequency interference. Due to packet birth
being a Poisson process, we can deduce that the signal interference ratio (SIR) is in neg-
ative correlation with node density. Therefore, we can describe node density with SIR.
Furthermore, we take SINR as the handoff criterion to eliminate the negative effects of
uneven density.

Definition 5.4: SINR is the ratio of received signal strength to the sum of Gauss white
noise and co-frequency interference.
From the SINR definition, we have

1 (5.24)
SINR =
1 1
+
SNR SIR

5.4.2.1.2  Handoff Decision

As well as keeping the signal quality, the VDL2 handoff will produce its overhead with
triggering logic-link establishment and executing data handover. The overhead will cause
the delay increase and capacity reduction in the process of handoff. To optimize the net-
work performance, the decision of handoff moment is the key point: being too dull will
worsen SNR and BER; being too sensitive will degrade system performance.
Taking the overhead into consideration, we applied a transition method in the handoff
decision. The method is different from the absolute decision (in which the handoff will be
triggered just after received signal strength is less than threshold) and supply a transition
interval (SQP_PARA) as the handoff trigger threshold instead of the absolute threshold, as
shown in Figure 5.13.
130 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

R2

SQP_PARA
SNR/dB
SN SNR3

SN
R1

Absolute
decision

Transition
decision
The moment to handoff

FIGURE 5.13
Threshold-based handoff decision.

The SINR-based handoff decision algorithm is as follows:

SQPnewGS ≥ SQPoriginGS + SQP_PARA (5.25)

5.4.2.2  Experiment Setup

To simulate the handoff process, we allocated 3 VDL2 GS and 16 aircraft with fixed flight
route in the simulation scenario, as shown in Figure 5.14.
In the scenario, we set the flight height to 10 km and the flight speed to 800 km/h. The
GS antenna height is set to 30 m.

5.4.2.2.1  Simulation Configuration

The handoff simulations are configured as follows:

1. Packet birth in any node meets Poisson distribution for which Poisson intensity is 5.
2. Packet length meets uniform distribution from 0 bit to 8192 bits.
3. P-CSMA access probability: GS side is 90/256 and AC side is 13/256.
4. The maximum transmission attempt number M1 is set to the default 135.
5. The DLS sublayer parameters are set by the defaults: the maximum number of
retransmissions N2 is set to 6; the length of sliding window is set to 4.

5.4.2.2.2  Performance Parameter

In practice, handoff will be triggered by one of the following conditions19:

Condition 1: The new GS signal quality is obviously larger than that of the origin GS.
Condition 2: N2 timer is expired, where system will believe the connection invalid.
Condition 3: Since AC did not receive GS transmitted packet over a long time (default
is 60 min), TG2 timer is expired.
 VDL2 Key Technology and Simulation

FIGURE 5.14
(See color insert.) Handoff scenario.
131
132 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Among the three conditions, condition 1 is the active mode, working in the normal com-
munication; conditions 2 and 3 are passive mode, working in the circumstance that can-
not keep normal communication and should be avoided. Therefore, the first demand is to
increase the active mode ratio and to guarantee the communication working in a normal
state. We will define it as the first-class performance parameter. Besides, we will still con-
sider the impact of handoff on overall performance and take it as the second-class perfor-
mance parameter.

1. The first-class performance parameter

Definition 5.5: Handoff ratio is the ratio of the passive handoff number to the
general handoff number.
2. The second-class performance parameter
Definition 5.6: Throughput is the ratio of successfully delivered bit rate and
channel capacity.
Definition 5.7: Delay is the time that a packet waits for before it is successfully
delivered.

5.4.2.3  Simulation Results

The section will simulate the performance of the SINR-based transition handoff, verifying
its feasibility and superiority.

5.4.2.3.1  Contrast Simulation

The purpose of this experiment is to verify if the proposed method can achieve obvious
improvement compared to the SNR-based absolute decision method. We choose the first
class as the performance parameter, which can reflect the handoff fundamental performance.
To make the contrast simulation, we modeled the two methods in VME module and then
made the contrast simulations in the same scenario as Figure 5.14. We also added debug-
ging statements in modeling so that we can observe the simulation objects such as target
GS, handoff type, and signaling exchange from OPNET debugger (ODB) window.
At the beginning of contrast simulations, the link establishment will be triggered for
all aircraft. Since the aircraft are well distributed at the initial time, we can assume that
they have the same interference and the link establishment depends only on SNR, which
is decided by the distance between AC and GS. The connection information (from ODB)
shows that ACs of both decision methods are similar at the initial linking and tend to link
to the GSs with the minimum distance.
With the developing simulation, the aircraft will fly along their respective tracks, and
the node density will change continuously. We can observe that the handoff due to dis-
tance (between AC and GS) will undergo in both methods. The difference exists in that the
SINR-based method will initiate handoff by excessive intensive density, but the SNR-based
method will not.
The contrast curves of handoff ratio are shown in Figure 5.15. In the figure, SINR-based
handoff curve is concave–convex with SQP_PARA ≈ 12 as the inflection point. ODB trace
shows that with the increase in SQP_PARA, the SIR-induced handoff proportion will be
continually falling in the active handoff and will be less than 50% when the curve arrives
at the inflection point. At the same time, the number of SIR-induced handoffs is far higher
than that of SNR-induced handoffs. Therefore, in the concave part, handoff is mainly
VDL2 Key Technology and Simulation 133

6
5.766

5
Handoff ratio (%)

2 SINR based handoff

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
SQP_PARA (dB)

FIGURE 5.15
The SINR-based handoff versus signal strength-based handoff.

caused by the intensive density, and the frequency of active handoff is comparably higher;
in the convex part, handoff is mainly caused by the distance (between AC and GS), and
the frequency of active handoff is comparably lower. On the other side, the active handoff
reduction will impair the system ability to allocate the radio resources and thus increase
the passive handoff ratio.
On the whole, handoff ratio increases with the increasing of SQP_PARA. In a wide area
(0–28 dB) that contains the inflection point, the SINR-based handoff rate is better than the
SNR-based handoff rate for which the optimum value is 5.766 as shown in Figure 5.15. The
improvement is derived from intensive density-based handoff trigger, which makes the
handoff being adaptive. When SQP_PARA exceeds 28 dB, the SINR-based handoff ratio
will continue to grow slowly, and handoff ratio tends to further deteriorate.
It should be pointed out that the handoff ratio curve will have a similar shape when the
aircraft number or route is changed. But with the number increasing, SIR-based handoff
will increase, and the inflection point will shift behind.

5.4.2.3.2  Transition Interval

This experiment will analyze the effect of SQP_PARA parameter on the handoff per-
formance and optimize the threshold. We choose the second class as the performance
parameter, which can reflect the handoff system performance. In a specified scenario, we
deployed 24 group simulations taking SQP_PARA as the variable, and each group lasted
60 min. To eliminate the accidental factors of flight route, we made 50 scenarios with dif-
ferent flight route schemes. The average throughput and delay can be obtained from the
multiple scenario-based simulations.
As shown in Figure 5.16, the average throughput curve has a quasi-parabolic shape with
downward opening. The average throughput will first increase, then decrease with SQP_
PARA growing up. It will reach the maximum value when SQP_PARA is about 8 dB. On
the contrary (as shown in Figure 5.17), the average delay curve has a quasi-parabolic shape
134 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

The result under a scenario

0.210
Average result under 50 scenarios

0.205

0.200
Throughput (%)

0.195

0.190

0.185

0.180

0 2 4 6 8 10 12 14 16 18 20 22 24
SQP_PARA (dB)

FIGURE 5.16
Throughput versus SINR threshold.

0.0154
The result under a scenario
Average result under 50 scenarios
0.0153

0.0152
Delay (s)

0.0151

0.0150

0.0149

0.0148
0 2 4 6 8 10 12 14 16 18 20 22 24
SQP_PARA (dB)

FIGURE 5.17
Delay versus SINR threshold.

with upward opening. The average delay will first decrease, then increase and will also
reach the minimum value at SQP_PARA ≈ 8 dB.
In this experiment, throughput and delay reflect the good consistency, namely, they can
simultaneously reach the optimal values at a point. The point is the optimal handoff thresh-
old, denoted as SQP_PARAopt. At the same time, the experiment shows that SQP_PARAopt
VDL2 Key Technology and Simulation 135

always falls in the area in front of the inflection point. The result indicates that SIR-based
handoff is dominant at the optimal point.

5.5 Conclusion
The chapter first discusses the application prospect of the VDL2 system, stating the
technical advantage of the VDL over the current ACARS and the necessity to realize
the system transition. Then, the four VDL modes are outlined. By comparison, the
conclusion is drawn that the VDL2 will dominate the air–ground VHF data link for
the ATN.
Focusing on the VDL2, the three-layered architecture is introduced on the whole, and
it is developed in detail by physical layer, data link layer, and subnetwork layer. We still
analyzed the key technologies in the VDL2 realization, involving D8PSK modulation, RS
channel encoding, P-CSMA, AVLC, handoff, etc. On the basis of the protocol and technol-
ogy, an entire VDL2 simulation model is built on the OPNET platform.
Last but not least, a handoff decision method is put forward based on SINR from the
view of co-interference and handoff overhead. To testify its superiority and feasibility,
a VDL2 handoff model is established on OPNET taking dynamic signaling interaction,
handoff timing, mobile node, multifrequency access, etc., as the modeling mechanisms.
Simulation experiments prove that the proposed method can realize handoff according to
the flight nodes density and reduce the passive handoff ratio effectively. Further simula-
tions also show that there exists optimal handoff threshold that can make the performance
such as throughput and delay to be optimized, which supplies a way to improve the cur-
rent handoff method.

References
1. Hui-yuan Z, Xue-jun Z, Jun Z. Migration from ACARS to new VHF data link in China.
Proceedings of International Conference on ITS Telecommunications, Chengdu, China, 2006:
pp. 1297–1300.
2. Kitaori J. A performance comparison between VDL mode 2 and VHF ACARS by proto-
col simulator. Proceedings of 28th Digital Avionics Systems Conference, Orlando, FL, 2009:
pp. 4.B.3-1–4.B.3-8.
3. ICAO. Annex 10 aeronautical telecommunications. Digital data communication systems.
Montreal, Quebec, Canada: ICAO, 2007.
4. James W R. Are we nearer to NEXCOM? Avionics Today (December 1, 2003). http://www.
aviationtoday.com/av/commercial/Are-We-Nearer-to-NEXCOM_1201.html#.VQku_tKl_YU.
Accessed: August 13, 2014.
5. RTCA. Signal-in-space minimum aviation system performance standards (MASPS) for advanced
VHF digital data communications including compatibility with digital voice techniques, RTCA
Standard DO-224A. Washington, DC: RTCA, 2000.
6. Bretmersky SC, Murawski RW, Konangi VK. Characteristics and capacity of VDL mode 2, 3,
and 4 subnetworks. Journal of Aerospace Computing, Information and Communication, 2005, 2(11):
470–489.
136 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7. ICAO. VHF Digital Link (VDL) (Mode 2) Standards and recommended practices,
AMCP/4-WP/70. Montreal, Quebec, Canada: ICAO, 1996.
8. ICAO, Annex 10. VDL-2 Standards and recommended practices. Montreal, Quebec, Canada:
ICAO, 1997.
9. Sklar B. Digital Communications-Fundamentals and Applications. Beijing, China: Electronic
Industry Press, 2002.
10. Xuehong C, Zongcheng Z. Information Theory and Coding. Beijing, China: Tsinghua University
Press, 2004: pp. 145–147.
11. Lin G, Zhi-jun W. A new method to analyze asynchronous P-CSMA. Journal of Computers, 2014,
9(1): 161–167.
12. ISO/IEC. Information technology—Telecommunications and information exchange between
systems—High-level data link control procedures, IEC 13239. ISO, 2002.
13. Whyman AW, Adnams MG. Proposed guidance material in support of VDL handoffs, AMCP
WG M. Malmo, Sweden: AMCP, 2000.
14. Lin, G, Zhi-jun, W. Performance analysis of MAC in VDL2 based on OPNET simulation. Applied
Scientific Research and Engineering Development for Industry. Zurich-Durnten, Switzerland: Trans
Tech Publications, 2013: pp. 1643–1646.
15. Li X, Ye M. Network Modeling and Simulation with OPNET Modeler, Xian Electronic Science and
Technology University Press: Xian, China, 2006: pp. 164–169, 212–214.
16. Jiasheng W, Jiahun, L. Stochastic Process. Tianjin, China: Tianjin University Press, 2003:
pp. 26–31.
17. Bretmersky S C, Konangi V K, Kerczewski R J. Performance of VDL mode 2 for the Aeronauti-
cal Telecommunications Network. Proceedings of 21st Digital Avionics Systems Conference, Los
Alamitos, CA, 2002: pp. 3.C.5.1–3.C.5.9.
18. Lin G. Key technology research of VDL2 aeronautical data link. Tianjin, China: Tianjin
University, 2014: pp. 90–96.
19. Desperier B, Esposito M, Delhaise P. VDL mode 2 hand-off analysis and simulation, AEEC DLK
Systems Subcommittee. Tampa, FL: AEEC, 2004.
6
GNSS Multipath Interference and Mitigation
for UAVs in Urban Canyon Environments

Saeed M. Khan

CONTENTS
6.1 Introduction......................................................................................................................... 137
6.2 Multipath Phenomenon..................................................................................................... 138
6.3 Evolution of GPS and Interoperability with the Galileo System................................. 140
6.3.1 Modernization and New Civil Signals for GPS................................................. 140
6.3.2 Impact of Evolution and Modernization on Urban Canyon–Type
Environments.......................................................................................................... 140
6.4 Establishing the Characteristics of an Antenna for Urban Canyon............................ 143
6.4.1 Choke Ring Antenna.............................................................................................. 143
6.4.2 Frequency Selective Surfaces (FSS) or Photonic Band Gap (PGB) Surfaces....... 144
6.4.3 Resistive Ground Planes........................................................................................ 145
6.4.4 Reduced Surface Wave Antennas......................................................................... 145
6.4.5 Conical Ring Antenna............................................................................................ 145
6.5 Comparisons of Four Related Designs............................................................................ 147
6.5.1 Impact of the Finite Ground Plane in the Absence of a Conical Ring............ 149
6.5.2 Effectiveness of the Conical Ring......................................................................... 149
6.5.3 Impact of FSS Elements.......................................................................................... 150
6.5.4 3-D RHC and LHC Pattern of Conical Rings...................................................... 151
6.6 Future Work......................................................................................................................... 152
6.7 Summary and Conclusions............................................................................................... 154
Bibliography................................................................................................................................. 157

6.1 Introduction
Unmanned aerial vehicles (UAVs) flying through urban canyons are subject to multipath
signals as a result of reflection, refraction, and diffraction of satellite signals traveling to the
receiver; this can lead to positional inaccuracies in this time-critical operation that can lead
to crashes. Indeed, analytical models show that for a GPS L1-C/A (coarse or clear signal/
acquisition), the time to detect a satellite signal increases considerably under conditions
of multipath interference (Schmid and Neubauer 2004, 503–509; Heng et al. 2014; Xie and
Petovello 2015). Traditionally, two different approaches have been taken to dealing with
the multipath problem. One approach involves signal processing and the other involves
designing an antenna (spatial processing) capable of rejecting multipath signals (Rahmat-
Samii 2003, 265–275; Nedic 2009; Kos et al. 2010, 399–402). While several signal processing

137
138 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

solutions for dealing with multipath impact exist, they require advanced receiver architec-
tures that either mitigate the effect of multipath signals or use signals identified as mul-
tipath to enhance location accuracy. A question that remains to be answered is whether or
not acquisition time for satellite signals can be optimized by carefully matching antenna
and receiver.
The purpose here is to suggest an antenna–receiver combination that is designed to cut
out multipath signals and to rely on line of sight (LOS) signals for location information.
Even with less than four LOS satellite signals in any one global navigation satellite system
(GNSS), the mobile unit can be located by using available receiver architectures that sup-
port multiple satellite navigation systems, for example, combining GPS and Galileo LOS
signals to determine position. The proposed solution employs an antenna design that
rejects multipath by cutting off signals that arrive at elevation angles low to the horizon; it
works with different receiver architectures and does not require additional processing time
for the purpose of multipath mitigation. Preliminary analysis has shown that the conical
ring antenna can outperform most antenna solutions in dealing with the multipath prob-
lem. Some common solutions rely on the use of choke rings (Basilio et al. 2005, 233–236),
frequency selective surfaces (FSS) or photonic band gap (PGB) surfaces (Rahmat-Samii
2003, 265–275), resistive ground planes, and the reduced surface wave (RSW) antennas
(Basilio et al. 2005, 233–236). In addition, the ring structure expected to provide protection
from jamming and spoofing signals incoming at angles low to the horizon.
It is expected that this treatment will provide the readers with a general physical under-
standing of the multipath phenomena and its importance to unmanned systems. The chapter
begins with a general introduction to the multipath environment that impacts both naviga-
tion and communication of unmanned aircraft systems (UAS) in an urban canyon setting.
The work will target both practitioners and researchers interested in reviewing the funda-
mentals of interference and important considerations for multipath mitigation. Following
the review, the special case of GNSS multipath interference in urban canyon environments
will be provided starting with background information about the evolution of GPS and its
interoperability with the Galileo system under the conditions of continued modernization.
Next, the author will present the results of his own research to a spatial mitigation tech-
nique that capitalizes on the multiple GNSS systems to cut down on the negative impact of
multipath. It is expected that this treatment will provide the readers with a general physical
understanding of the multipath phenomena and its importance to unmanned systems.

6.2  Multipath Phenomenon

Although radio frequency interference, multipath, and ionospheric scintillation are all
potential sources of error, technological advancements in GNSS systems have diminished
the impact of most sources of error to the point where multipath and shadowing (a phe-
nomenon where the direct path suffers excess attenuation) remain as significant issues
(Ward et al. 2006, 279–295). The multipath phenomenon is experienced by a receiver when
it receives either a direct path signal from a source along with one or more reflected/
diffracted satellite signals. If one visualizes these signals to be following the path of the
rays in Figure 6.1, then the receiver is being subjected to multipath signals that can lead to
errors. In this simplified figure, multipath signals arrive at the receiver through reflected
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 139

High-elevation satellite source

Direct

Reflected
multipath

Diffracted
multipath
Low-elevation
UAV satellite source

Reflected Direct path

multipath with shadowing

FIGURE 6.1
Direct/multipath headed to a receiver on a car and a UAV. Only one satellite signal is used in each case.

and diffracted angles to the UAV from a single high-elevation satellite source. A low-
elevation satellite source signal is received at the car by both a multipath reflected signal
and a direct signal that has undergone shadowing while passing through foliage. The
strength of the shadowed direct signal is attenuated and in some cases may even be lower
in level than the multipath signal and this can cause problems in discrimination between
the signals at the receiver. It should be noted that we have ignored the compound effect of
all satellites in view that should be affecting both receivers.
In order to understand the impact of multipath on receivers, it is important to review how
distances to satellites are measured. Distances to the satellite vehicle (SV) from receiver can
be measured by cross-correlating the transmitted pseudorandom (PRN) code to a known
PRN for the particular SV at the receiver until the phases of both codes match. If the clocks
on the SV and the receiver can be synchronized, then the time shift Δt required to match
phases can be multiplied with the speed of light to gives us the distance (c·Δt). Errors are
introduced in the calculation of this time shift in a multipath environment, and this can
lead to positional inaccuracies.
Not all multipath signals introduce errors. Multipath signals travel different path
lengths to get to the receiver and can be broken down into near echoes and far echoes.
If the time of arrival between two echoes is greater than the inverse of the system band-
width, then they can be resolved (Jahn et al. 1996, 1221–1226). Therefore, multipath delays
are greater than twice the spreading code period they are resolvable (Ward et  al. 2006,
279–295), and as such multipath from nearby objects (near echoes) can distort the correla-
tion functions. For example, if a receiver uses C/A-code at a chipping rate of 1.023 MHz
multipath with path lengths greater than 600 m longer, then the direct signal path can
be resolved. Problems can also stem from grazing reflected signals arriving from greater
distances as they tend to destructively interfere with the direct signal by being in phase
opposition. The impacts of different ray paths are summarized in Table 6.1.
140 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 6.1
Ray Path and Impact
Ray Path Distance/Travel Time Path Characteristics Impact Comment
Direct path to receiver Direct with no Unobstructed Does not introduce multipath
w/o shadowing attenuation in channel errors.
Direct path to receiver Direct with attenuation Obstructed by objects Excessive attenuation can cause
with shadowing in channel that cause signal strength fall below
attenuation in signal multipath signals causing
resolving issues.
Multipath near Arrive within twice the Reflected/diffracted Distorts the correlation function
echoes spreading code period one or more times received composite signal and
after direct signal locally generated reference.
Multipath far echoes Arrives taking more than Reflected/diffracted Resolvable from near echoes in
twice the time for the one or more times receiver.
spreading code period
following direct signal
Grazing reflected rays Almost the same time Reflected once In phase opposition to direct
At low angles to the taken as the direct path path; can potentially cancel
horizon out the direct path ray.
Source: Ward, P. et al., Interference, multipath, and scintillation, in Understanding GPS Principles and Applications,
2nd edn., E. Kaplan and C. Hegarty (eds.), Artech House, Norwood, MA, 2006, pp. 279–295.

6.3  Evolution of GPS and Interoperability with the Galileo System

Since the 2004 GPS–Galileo agreement establishing cooperation between the U.S. and EU
member states, eight major documents describing the path and course of this cooperation
have come out. Table 6.2 is a summary of these documents and describes the historical pro-
gression of this collaboration. It is fairly easy to see that the path forward includes, among
other things, optimized civil signals that would allow future receivers to track GPS and/
or Galileo systems with higher accuracy. The U.S. government has also started cooperation
conversations with India, China (most recent), Japan, Australia, and Russia (currently on
hold).

6.3.1  Modernization and New Civil Signals for GPS

Another bit of good news for civilian users is that the U.S. government has started to bring
into the picture three new signals for their use, which includes the L2C, L5, and L1C. The
L1C capabilities will start to come online in 2016 with GPS III phase and be available on
24 satellites around 2026 (U.S. Air Force 2014b). Modernization plans for both GPS and
Galileo will make it possible for upgraded receivers to track L1C GPS and/or E1 Open
Service Galileo signals with higher accuracy. The following table summarizes the modern-
ization status of these new civil signals in Table 6.3.

6.3.2  Impact of Evolution and Modernization on Urban Canyon–Type Environments

The previous discussion of modernization and evolution of GNSS systems presents some
new opportunities for the spatial processing of multipath signals in an urban canyon.
Following is an itemized list of some of these points:
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 141

TABLE 6.2
Historical Progression of the GPS–Galileo Cooperation
Year and Document Key Points
2004 GPS-Galileo Agreement Framework for cooperation between the Parties in the promotion, provision
(U.S. Air Force 2004) and use of civil GPS and GALILEO navigation and timing signals and
services, value-added services, augmentations, and global navigation and
timing goods. The Parties intend to work together, both bilaterally and in
multilateral fora, as provided herein, to promote and facilitate the use of
these signal, services, and equipment for peaceful civil, commercial, and
scientific uses, consistent with and in furtherance of mutual security
interests.
2006 Joint Statement on EC and the United States signed the Agreement on the Promotion, Provision
Galileo-GPS Signal and use of Galileo and GPS Satellite-Based Navigation Systems and Related
Optimization (U.S. Applications on the compatibility and interoperability of the Galileo and
Air Force 2006) GPS satellite navigation systems. A central element of the Agreement was a
common baseline signal structure that could be optimized for greater
performance.
2007 Joint Statement on Agreement on the Promotion, Provision and Use of Galileo and GPS
GPS-Galileo Working Satellite-Based Navigation Systems and Related Applications. The purpose
Group B (U.S. Air Force of Working Group B is to consider, inter alia, non-discrimination and other
2007) trade related issues concerning civil satellite-based navigation and timing
signals or services, augmentations, value-added services, and global
navigation and timing goods.
2007 Joint Fact Sheet on The United States and the European Union are designing GPS and Galileo to
GPS-Galileo Cooperation transmit one or more common civil signals, so future users will enjoy the
(U.S. Air Force 2007) benefits of multiple PNT satellite constellations. Benefits include increased
satellite availability (particularly in urban environments) and improved
resistance to signal interference. Signals will be provided on two common
frequencies by the United States and the European Union without direct
user fees.
2007 Joint Press Release on The resulting GPS L1C signal and Galileo L1F signal have been optimized to
Civil Signal Design (U.S. use a multiplexed binary offset carrier (MBOC) waveform. Future receivers
Air Force 2007) using the MBOC signal should be able to track the GPS and/or Galileo
signals with higher accuracy in challenging environments that include
multipath, noise, and interference.
2008 GPS-Galileo Working The Working Group addressed specific questions about Galileo raised by U.S.
Group B Meeting (U.S. industry and considered proposals for enabling global equipment
Air Force 2008) manufacturers to develop and test GPS/Galileo receivers and applications
on a level playing field that benefits users.
2008 Joint Statement first The improved common civil signal, referred to as L1C on GPS and E1 Open
plenary meeting (U.S. Service on Galileo, has been optimized using a Multiplexed Binary Offset
Air Force 2008) Carrier (MBOC) waveform. Future receivers using this signal should be able
to track the GPS and/or Galileo signals with higher accuracy in challenging
environments.
2010 Joint Statement on The combination of GPS and Galileo services provided noteworthy
Combined Performance of performance improvements particularly in partially obscured environments,
GPS and Galileo (U.S. where buildings, trees or terrain block large portions of the sky. Dual-
Air Force 2010) frequency receivers provide additional improvements in most environments.
2013 Interim Report of Working Group C (WG-C) is designed to enhance cooperation for next
ARAIM Technical Subgroup generation GNSS. One of the objectives of WG-C is to develop integrated
(U.S. Air Force 2013) applications for Safety-of-Life services. To this end, WG-C established the
ARAIM Technical Subgroup. The objective of the Subgroup is to investigate
ARAIM (Advanced Receiver Autonomous Integrity Monitoring). The
further goal is to determine whether ARAIM can be the basis for a
multi-constellation concept to support air navigation worldwide.
142 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 6.3
New Civil Signals
Frequency and
Signal Modulation Launching Dates Key Feature
L2C (second GPS 1227.6 MHz; Started launching 2005 and has First to provide dual frequency
civilian signal) biphase shift key 14 GPS satellites as of October capabilities to civil signals,
(BPSK) operation 7, 2014; 24 GPS satellites leading to higher-precision
available around 2018 locational capabilities than
with the L1 C/A alone
L5 (third GPS 1176.45 MHz; Started launching 2010 and has Greater bandwidth for
civilian signal) biphase shift key 7 GPS satellites as of October 7, improved jam resistance;
(BPSK) operation 2014; 24 GPS satellites higher transmitted power
available around 2021 than L1 C/A and L2C
L1C (fourth GPS 1575.42 MHz; Begins launching with GPS III Designed for international
civilian signal) Multiplexed Binary in 2016 moving to 24 satellites GNSS interoperability
Offset Carrier with capability in 2026
Modulation
Source: U.S. Air Force, New civil signals, 2014a, http://www.gps.gov/systems/gps/modernization/
civilsignals/#L1C, accessed November 9, 2014.

High-elevation satellite sources

from different GNSS systems

Sufficient direct ray paths

possible with multisystem
receiver

UAV

FIGURE 6.2
Multisystem receiver would benefit from more overhead SVs in an urban canyon.
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 143

1. The development and maturing of more and more GNSS (Table 6.2) systems
over time will increase the focus on developing multisystem receivers allowing
receivers to take benefit of more unobstructed direct line signals in an urban
canyon-like environment (Figure 6.2).
2. The new developments bring up prospects for a spatial processing system that cut
of all signals from SVs that are not at high elevation.
3. The coming on line of new civilian signals L2C, L5, and L1C will allow higher
levels of precision (Table 6.3) even if low-elevation satellites are cut out, since
additional signals would be available and at different frequencies.
4. The cutting off the low-elevation SVs will cut off grazing reflected rays and near
echoes at these angles as well (see Table 6.1).
5. Based on the trajectories and cooperation between countries in the area of GNSS,
there will be increased focus on the design and development on multisystem
receivers.
6. Based on items 1–5, one can see that the demand for spatial processing anten-
nas for use in urban canyon-like environments will increase and also will be
supported by an availability of new multisystem receivers.

6.4  Establishing the Characteristics of an Antenna for Urban Canyon

GNSS antenna design considerations that respond to the need for urban canyon-like envi-
ronments can be established by using the evolution and modernization trajectories along-
side important electromagnetic considerations of good bandwidth (especially important
when using multiple systems), axial ratio, and radiation pattern. To begin our discussion,
we focus on some different multipath-rejecting antennas that are quite well known and the
desirable characteristics that they exhibit but are not necessarily ideal for UAV application.

6.4.1  Choke Ring Antenna

While the choke ring antennas have been primarily used for precision in measuring geo-
detic references, it is has established itself as a well-known survey antenna and has many
known variations (Basilio et al. 2005, 233–236; Sciré-Scappuzzo and Makarov 2009, 33–45).
The basic structure of the antenna consists of a number of concentric metallic cylinders
that sit on top of a ground plane, creating a corrugated surface (Figure 6.3 illustrates the
basic structure of a choke ring antenna). The surface prevents surface waves and plane
waves from creating edge radiation and thereby reduces backlobe radiation and radiation
at low elevation angles from the horizon. This antenna also has good phase center (the
apparent point from which the electromagnetic wave spreads spherically; it should be
noted that this is not a single point but varies in position based on the angle of incidence)
characteristics, that is, small variation with the angle of incidence. The depth of the corru-
gation should be greater than quarter wavelength of its lower operational frequency and
less than half the wavelength of its upper operating frequency (about 2.5 in. for an L1/
L2 antenna). The size of the ground plane is also an important consideration (about 15 in.
for an L1/L2 antenna). The size and weight of this antenna make it an unlikely choice for
an UAV.
144 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Prevents signals
low angles to horizon

Side view

Ground plane

Concentric rings Antenna

element

Top view

FIGURE 6.3
Choke ring antenna.

6.4.2  Frequency Selective Surfaces (FSS) or Photonic Band Gap (PGB) Surfaces
FSS or PGB surfaces when used as ground planes have been used to cut down on sur-
face waves and edge radiation by offering high impedances to surface currents on them.
This high impedance property works only within a design bandwidth (Sievenpiper et al.
1999, 2059–2073).
The goal of FSS is to come up with smaller, thinner, and lighter ground planes for anten-
nas that broaden the possibility of deployment in many more applications when compared
with survey antennas like the choke ring antenna. Generally, these FSS are constructed by
placing of elements on top of substrates spread out like an array. While many different ele-
ment structures exist, such as variations of square loops and split rings, these elements are
carefully designed such that their interaction with the ground plane and with each other
produces the proper capacitive gaps and inductive traces to create resonances (high imped-
ance) to surface currents. The elements can be as small as λ/12 (Bayatpur and Sarabandi
2007, 1239–1245; Bayatpur 2009, 1–164) and are not considered to be sensitive to the angle of
incidence. The elements can be placed very close to the antenna and this can help reduce
the dimensions of an optimal design (Figure 6.4 shows an FSS surface with antenna).

Antenna element

Split ring
elements laid
out in array Array of elements
prevents surface waves
and edge radiation

FIGURE 6.4
FSS and antenna.
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 145

6.4.3  Resistive Ground Planes

Another method of shaping the radiation pattern to avoid multipath is to use resistive
cards at the edges of the antenna ground plane (Rojas et al. 1995, 1224–1227). The resistive
edges (Figure 6.5) of the ground plane prevent diffraction and help shape the radia-
tion pattern such that it offers less gain to signals arriving at low angles to the horizon.
The resistive edges, helps prevent distortion of the phase center due to this edge effect.

6.4.4  Reduced Surface Wave Antennas

RSW antennas have a promising design for rejecting multipath (Basilio et al. 2005, 233–236).
The shorted-annular-ring-reduced surface wave antenna consists of a circular patch that
shorted to the ground plane at a carefully selected distance from the center that is based on
the analysis required to prevent lateral radiation. The performance of this antenna when
sitting on a 35.6 cm ground plane proved to be promising when compared to the choke
ring antenna pattern gain on the horizon (10 dB lower and better at cutting off multipath
arriving at angles low to the horizon). This antenna would prove to be at least an order of
magnitude lighter than the choke ring.

6.4.5  Conical Ring Antenna

The conical ring antenna (Khan 2014, 232) has been designed especially for UAVs in urban
canyon following the design criteria listed as follows:

1. Design should compare favorably or exceed in multipath rejecting capabilities of

choke ring antennas; antennas using FSS or PGB surfaces; antennas using resistive
edges on their ground planes; and RSW antennas.

Side view Edges help prevent

diffractive effects

Resistive edges

Antenna

Ground plane

Top view

FIGURE 6.5
Antennas with resistive edges on their ground planes.
146 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

2. Should be able operate at or be extendable to both L1/L2 frequencies. Such that it

can take advantage of the L1C/L2C and the E1 Open service for Galileo.
3. It should have the size, weight, and shape that make it a deployable payload for
both small and large UAVs.

The structure of the conical ring antenna is shown in the Figure 6.6. The lowest layer of
the antenna is a feed structure for this multifeed antenna. The feed structure is respon-
sible for creating four equal amplitude excitations that are phased 0°, 90°, 180°, and 270°
to produce right circularly polarized radiation (RHCP). Four probes carrying these exci-
tations emerge from below a metallic ground plane and pass through ground plane
below the patches without touching either one. The feeding posts also serve to mechani-
cally support the ring and patch combination. Once it passes through the substrate of
the patch, it makes contact with the top patch. In the event a dual frequency (say L1/L2)
design is used, the probes would pass through the bottom patch without making con-
tact with the top patch. The conical ring is supported by a dielectric structure built by
a 3-D printer.
The model conical ring is backed by a dielectric material called abs plastic (relative
dielectric constant 2.87). Other dielectric materials (polyethylene, fiber glass, ceramics,
etc.) can also be used, provided the dimensions of other components are scaled properly.

Circular patch
or stacked patches

Conical ring
(conducting)
Dielectric support
for conical ring Feed probes

Feed network
providing properly
phased excitation
Side view slice
Patch/patches

Ground plane

Conical
ring

Top view

FIGURE 6.6
Construction of conical ring antenna (not to scale).
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 147

TABLE 6.4
Gain Comparison with Different Solutions at Elevation Angles Close to the Horizon
Angle in Degree 90 85 80 75 70
RSW (measured (Basilio et al. 2005, 233–236)) −30 dB −25 dB −21 dB −15 dB −11 dB
Conical ring (simulated) −31 dB −29.5 dB −25 dB −21 dB −17.5 dB
Choke Ring (measured (Basilio et al. 2005, 233–236)) −18 dB −16 dB −15 dB −13 dB −10 dB
A smaller value of gain, at these elevation angles, indicates better multipath rejection.

Currently, the ground plane is about 16 cm in diameter, post height is about 1 cm above
ground plane, circular patch has a diameter of about 7.2 cm. Increasing the dimensions
of either the conical ring and/or circular patch inside will lower the resonant frequency.
Using a material of higher dielectric constant will also lower resonant frequency.
When compared with choke rings (Basilio et al. 2005, 233–236), the conical antenna dem-
onstrates superior multipath rejection in all cases of elevation angles between 70° and 90°
(or angles low to the horizon) as evidenced by the pattern gain. The conical ring gains were
between 8 and 13 dB lower for the field pattern. The largest dimension of conical antenna
without a radome is about 16 cm whereas the largest dimension for choke rings is around
30 cm. There will also be a considerable difference in weight which makes the conical ring
also more desirable from a payload point of view.
When the RSW antenna (Basilio et  al. 2005, 233–236) sits on a 35.6  cm ground plane,
the conical antenna demonstrates superior multipath rejection in all cases of elevation
angles between 70° and 90° (or angles low to the horizon) as evidenced by the pattern
gain. Regardless of the RSW antenna itself, for these results it was sitting on a large ground
plane more than four times the largest dimension of the conical ring. While the choke ring
might be an order or more heavier than the RSW plus ground plane combination, the coni-
cal ring beats them both in the desirable size and weight criteria.
While antennas with resistive ground planes have not been compared against the coni-
cal ring, the paper describing the E-plane pattern (Rojas et al. 1995, 1224–1227) in 70°–90°
elevation finds that resistive ground provides a minimum gain that is greater than −13 dB.
The best and worst case performance for the conical ring in this range is −31 dB (at 90°)
and −17.5 (at 70°) indicating superior performance in this case as well (Table 6.4).

6.5  Comparisons of Four Related Designs

In order to study how structural variations in conical ring antenna impact the antenna and
in order to conduct further analysis of its behavior, four different antenna structures were
compared. These structures include a circular patch, circular patch with FSS surface, the
conical ring antenna (which is simply a circular patch surrounded by a conical ring), and a
conical ring surrounded by an FSS surface. The dimension of the ground plane is kept the
same in all cases and each structure is fed using four feed signals having equal amplitudes
and phase to deliver.
To start, a simulation for the radiation pattern for all four are plotted for comparison
(Figure 6.7) and their multipath rejecting capabilities are compared (Table 6.5). From
Table 6.5, comparing a plain circular patch with the circular patch that has been sur-
rounded by FSS elements, it can be seen that the latter provides a better multipath rejecting
148 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS
Dashed line Dash dot line Solid Line Dash dot line

CONE/FSS/FSS + CONE/plain patch

0

–10
RHC on cut 1 (at 0°), RHC on cut 1 (at 0°),... (dB)

–20

–30

–40

–50

–60

–70

–80

–90
–180 –150 –120 –90 –60 –30 0 30 60 90 120 150 180
Angle (°)

FIGURE 6.7
(See color insert.) Patterns of all four structures compared.

TABLE 6.5
Gain Comparison with Different Solutions at Elevation Angles Close to the Horizon—RHC Pattern
Angle in Degree 90 85 80 75 70
Circular patch −12 dB −11 dB −10 dB −9 dB −8 dB
Circular patch with FSS surface −20 dB −18 dB −16 dB −14 dB −12 dB
Conical ring −31 dB −29.5 dB −25 dB −21 dB −17.5 dB
Conical ring with FSS −18.5 dB −21 dB −24 dB −32.5 dB −33 dB
A smaller value of gain, at these elevation angles, indicates better multipath rejection.
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 149

capability at close to horizon angles. This is evident from the table as the circular patch
with FSS elements provides 8 dB lower gain at 90° and 4 dB lower gains at 70°.
When a comparison is made between the conical ring and the circular patch with an
FSS surface, the conical ring shows superior multipath rejection capabilities by having
anywhere between 6 and 11 dB less gain. An interesting situation arises when the conical
ring is compared with the conical ring plus FSS. While the conical ring does outperform
the conical ring with FSS in the 70°–90° range, there is a range around 60° when the one
with FSS has better rejection. A question that this raises is whether or not it is possible to
tune these behavior-changing elements in the FSS array.

6.5.1  Impact of the Finite Ground Plane in the Absence of a Conical Ring
We know that for an ordinary circular patch, the radiation pattern, directive gain, and
input impedance vary widely with the size of the ground plane. Analytical studies
(Bhattacharyya 1990, 152–159) have shown that for a circular patch with a ground plane
radius of 0.63λ0 (where λ0 is the free space wavelength), the patch antenna achieves maxi-
mum gain. Large ground planes are undesirable for most applications, especially when
designed for small UAVs.

6.5.2  Effectiveness of the Conical Ring

With the inclusion of the conical ring (Khan 2014, 232), the RMS electrical field distribu-
tion, RMS magnetic field distribution, and the time average power density in the near
field (Figures 6.8 through 6.10), when compared to the plain patch, appear much more
symmetrical and tightly bounded to the phase center. (In practical terms, when referenced
to the phase center, the fields radiated by the antenna are spherical waves with equiphase
surfaces over most of the angular space in the main beam.) The phase center for the conical
ring antenna is located 1 mm above the ground plane as per simulation results; the field
and power distribution is from a slice taken at 7 mm above the ground plane.

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Electric (dB)

FIGURE 6.8
(See color insert.) RMS electric field distributions in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.
150 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Magnetic (dB)

FIGURE 6.9
(See color insert.) RMS magnetic field distributions in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Power density (dB)

FIGURE 6.10
(See color insert.) Time average power density in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.

In the previous sections multipath rejection benefits of the conical ring antenna were
noted. The tightly held near field should require a smaller ground plane making it
possible for the antenna dimensions to be further reduced. Analytical, numerical, and
experimental studies are needed to investigate the impact of the ring on size of ground
plane.

6.5.3  Impact of FSS Elements

Observations of field and power distributions for all four antennas from simulation
(Figures 6.7 through 6.10) indicate that the use of FSS surfaces in combination with the
conical ring has an interesting radiation pattern that allows a null to form further away
from the horizon (at a smaller elevation angle ϴ = ±70° horizon is at ϴ = ±90° elevation).
Figures 6.8 through 6.10 show how the circular patch with FSS elements surrounding
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 151

it creates a more symmetrical field and power distribution (with or without the coni-
cal structure) when compared with an ordinary circular patch. With a conical structure
added on the RMS electric field distribution seems to spread further toward the edges of
the ground plane.
Characteristics of FSS surfaces are almost entirely determined by FSS elements, which in
this case is the ring structure. One of the properties of resonant length standard FSS ele-
ments is its reflection and refraction depend on the angle of incidence (Note: This is not the
case for smaller nonresonant lengths). Therefore, by changing the FSS element structure
it should be possible to change the null location on their radiation patterns. This impact
needs to be further investigated as it provides an additional degree of freedom to control
where nulls form.

6.5.4  3-D RHC and LHC Pattern of Conical Rings

Figures 6.11 and 6.12 shows the 3-D RHC and LHC patterns, respectively. The RHC pattern
shows a maximum gain of about 9.3 dBi (decibels isotropic). The pattern also shows very
low gain for elevations that corresponds to low angles to the horizon and primary sources.
The symmetrical nature of the patterns is both a function of the symmetry of the antenna
and the progressively phased equal amplitude feeds. The 3-D LHC pattern also shows
very low gain for near horizon and this is a plus as well in the event that there is a change
in the handedness of the reflected signals.

RHC pattern

Y
ERHC

Directivity: 9.274 dBi Frequency: 1.575 GHz

Polrzn. loss: –57.604 udB
Material loss: Not known Radial scale: Linear in field magnitude
Mismatch: Not Known Ant eff: 100.000% Scale max at: 0 dB–directivity
Gain: 9.274 dBi Rad eff: 100.000% Contour at: –3 dB–directivity

FIGURE 6.11
(See color insert.) 3-D RHC radiation pattern.
152 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Comparing conical ring, FSS, and ordinary circular patch

Z

ELHC

Directivity: 9.709 dBi Frequency: 1.575 GHz

Polrzn. loss: –60.884 dB
Material loss: –18 .581 mdB Radial scale: Linear in field magnitude
Mismatch: Not known Ant eff: 99.573% Scale max at: 0 dB–directivity
Gain: –51.194 dBi Rad eff: 99.573% Contour at: –3 dB–directivity

FIGURE 6.12
(See color insert.) 3-D LHC radiation pattern.

6.6  Future Work

Some of the unresolved issues of the conical antenna are as follows:

1. While the dimension of the conical antenna is much smaller than other well-
known multipath rejecting types, and while it compares favorably with or exceeds
the capabilities of these antennas, its largest physical dimension at 16 cm is still
quite large for some smaller UAVs. Future work should explore the impact of using
a higher dielectric constant substrate and support materials and the ability to scale
down the size of the path, conical ring.
2. Although the system is designed to accommodate stacked patches allowing for
dual frequency operation (e.g., L1/L2), this has not been studied yet. Stacking to
patches on top of one another where the feed probe comes in without connect-
ing with the patch and then makes contact with the top patch is used in many
applications (Figure 6.13 illustrates this concept). The two patches are parasitically
coupled and provide two distinct resonances and can be tuned to the dual fre-
quencies being covered.
3. In Table 6.1, it was mentioned that not all multipath rays cause positional errors.
Rays that arrive after the direct ray by more than two chip code periods does
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 153

Hole in Probe
ground plane feed

Dielectric Top patch

Bottom view w/o feed cable
substrates
Bottom
patch

Through
hole

Outer conductor
of feeder cable

Section view

FIGURE 6.13
Stacked patches.

not create multipath errors since they can be resolved from multipath within this
limit. So nearby reflection is more important in this case. These reflections might
be coming from the aircraft itself or other aircraft flying in close formation. The
proper placement of the antenna on the aircraft body might solve this problem,
since we are now talking about signals coming from high angles as the low angles
are afforded protection by the ring structure. A reflected ray is likely to cause
a problem if it reflects onto the antenna from a higher point than the antenna.
Keeping the antenna properly spaced from those at higher points might avoid
multipath (Figure 6.14) from the aircraft body.
When it comes to close formation flights, the problem does not go away as eas-
ily as there is no guarantee that the reflections are not coming at high angles from
other aircraft (Figure 6.15). There are good reasons and opportunities to explore
further what happens in close formation and how to arrive at an ideal placement
of antennas on an aircraft.
4. Finally, for future work should test antenna in the multipath environments resem-
bling an urban canyon. As it is not possible to fly UAVs without a certificate of
authorization (COA) from the FAA, a ground vehicle can be used for testing.
Driving a truck through a metropolitan downtown with the test antenna hooked
onto a multisystem receiver (e.g., SX-NSR) while conducting GPS only, Galileo
only, and dual-tracking tests is critical, with and without the use of a hybrid
system. Figure 6.16 shows how these systems will be mounted on trucks. Note
that the antenna is actually mounted on an UAV and the combination is strapped
154 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

e
om Conical ring
Rad

Patches

Expanded
view of mounted
antenna on UAV

High angles
accepted
Reflected rays
from body

FIGURE 6.14
Antenna receiving high angle multipath signals from aircraft body.

onto a truck. In the event a COA is available, one can proceed with such a test in
an authorized area with the antenna and receiver as payload (Figure 6.17 shows an
UAV that might be able to carry such a payload).

6.7  Summary and Conclusions

This chapter discusses the key issues associated with GNSS positioning accuracy in an
urban canyon–type environment where multipath can lead to errors and errors can lead
to catastrophic situations for UAVs and their payloads. Spatial processing solutions to
multipath are being preferred here to ones that use signal processing. More importantly,
one particular spatial processing antenna is being forwarded, both at satisfying specifica-
tions and at establishing a design criteria that was used to come up with these specifica-
tions that is most urgently needed at this time given the evolutionary process at play in the
GNSS area.
The basic assumption here is that there are an increasing number of GNSS systems and
satellites coming into play with enough interoperability to allow someone to depend on
a multisystem approach to positioning, whereby they can cut signals arriving at small
angles to the horizon and rely only on overhead satellites to determine their position.
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 155

Reflected signal arriving

from a formation
aircraft

FIGURE 6.15
Antenna receiving high angle multipath for another formation aircraft.

Antenna under test

Unmanned
aerial vehicle

FIGURE 6.16
(See color insert.) A possible antenna test mount for urban canyon environment.
156 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

FIGURE 6.17
(See color insert.) A KSU Salina UAS asset. (Aerosonde Mark 4.7 UAS platform.)

Another key assumption is that by cutting off signals at angles low to the horizon one
can reduce multipath errors. Both these assumptions lead to this new design criterion
that, in terms of antenna speak, can be interpreted as having radiation patterns that
have lowest gains at or near the horizon. Figure 6.7 illustrates how well the conical ring
antenna performs in this area (it should go without saying that the old design criteria of
good axial ratio, impedance matching, bandwidth, high gain in desirable directions still
remain intact). Table 6.4 shows how well the conical ring performs at elevation angles
close to the horizon and outperforms both the RSW and choke ring antennas by the new
criteria.
The long-standing mechanical criteria of being a compact lightweight structure cannot
be overemphasized in this application. Here also the conical ring betters RSW and choke
ring in terms of size and weight. Figures 6.8 and 6.9 are significant in that it shows the RMS
electric and magnetic field distribution is tightly held in the center. This is a likely reason
why good roll off is achieved with a relatively small ground plane. Also from Figures 6.8
and 6.9 it is interesting to note how the FSS elements have impacted the RMS field distri-
bution. It seems that the FSS elements under the cone tend to spread the field distribu-
tion outward toward the edges of the ground plane and are therefore more prone to edge
effects.
Much of the discussion thus far has centered on the conical antenna’s role as an antenna
on a UAV. Given how the antenna rejects ray paths low to the horizon, it should provide
some protection from jamming and spoofing sources that are ground based as well.
It should be stated that no significant research has been done by the author with regards
to protection from spoofing and jamming by the conical ring and that more investigation
is needed.
Finally, studies are needed to see if the physical dimensions of the antenna can be
reduced further without unacceptable loss in performance. The interesting problem of
what happens when multipath rays actually enter the cone through near vertical angles
from the body of the same aircraft as well as other aircraft flying in close formation is
a real-world problem needing a solution. Finally, a real-world test where the antenna is
taken for ride in an urban canyon hooked up to a multisystem receiver is important for
gathering performance feedback.
GNSS Multipath Interference and Mitigation for UAVs in Urban Canyon Environments 157

Bibliography
Basilio, L., J. Williams, D. Jackson, and M. Khayat. 2005. A comparative study of new GPS reduced-
surface-wave antenna. IEEE Antennas and Wireless Propagation Letters 4: 233–236.
Bayatpur, F. and K. Sarabandi. 2007. Single-layer, high-order, miniaturized-element frequency
selective surfaces. IEEE Transactions on Microwave Theory and Techniques 55(5): 1239–1245.
Bayatpur, F. 2009. Metamaterial-inspired frequency-selective surfaces. PhD, University of Michigan,
Ann Arbor, MI.
Bhattacharyya, A.K. 1990. Effects of finite ground plane on the radiation characteristics of a circular
patch antenna. IEEE Transactions on Antennas and Propagation 38(2): 152–159.
Heng, L., T. Walter, and G. Gao. 2014. GNSS multipath and jamming mitigation using high-mask-
angle antennas. IEEE Transactions on Intelligent Transportation Systems 16: 741–750.
Jahn, A., H. Bischl, and G. Heiβ. 1996. Channel characterisation for spread spectrum satellite com-
munications. Proceedings of IEEE Fourth International Symposium on Spread Spectrum Techniques
and Applications (ISSSTA’96), Mainz, Germany, pp. 1221–1226.
Khan, S. 2014. A comparative analysis of the impact of different structures on multipath mitigation
in the GPS context. 2014 IEEE International Symposium on Antennas and Propagation and USNC-
URSI Radio Meeting, Memphis, TN.
Kos, T., I. Markezic, and J. Pokrajic. 2010. Effects of multipath reception on GPS positioning perfor-
mance. 2010 Proceedings ELMAR, Zadar, Croatia, pp. 399–402.
Nedic, S. 2009. On GPS signal multipath modeling in dynamic environment. 2009 Aerospace Conference,
Big Sky, MT.
Rahmat-Samii, Y. 2003. The marvels of electromagnetic band gap (EBG) structures: Novel microwave
and optical applications. Proceedings of the 2003 SBMO/IEEE MTT-S International 1: 265–275.
Rojas, R.G., D. Colak, and W.D. Burnside. 1995. Synthesis of tapered resistive ground plane for
resistive antenna. 1995 Antennas and Propagation Society International Symposium, Newport
Beach, CA.
Schmid, A. and A. Neubauer. 2004. Comparisons of sensitivity limits for GPS and Galileo receivers
in multipath scenarios. Position Location and Navigation Symposium, Monterey, CA, pp. 503–509.
Sciré-Scappuzzo, F. and Sergey Makarov. 2009. A low-multipath wideband GPS antenna with cutoff or
non-cutoff corrugated ground plane. IEEE Transactions on Antennas and Propagation 57(1): 33–45.
Sievenpiper, D., L. Zhang, R.F.J. Broas, N. Alexópolous, and E. Yablonovitch. 1999. High-impedance
electromagnetic surfaces with a forbidden frequency band. IEEE Transactions on Microwave
Theory and Techniques 47(11): 2059–2073.
U.S. Air Force. 2004. 2004 Agreement on the promotion, provision and use of Galileo and GPS
satellite-based navigation systems and related applications. Accessed November 9, 2014.
­
http://www.gps.gov/policy/cooperation/europe/2004/gps-galileo-agreement.pdf.
U.S. Air Force. 2006. 2006 Joint statement on GPS-Galileo signal optimization.
Accessed November 9, 2014. http://www.gps.gov/policy/cooperation/europe/2006/​joint-​
statement/.
U.S. Air Force. 2007a. 2007 Joint fact sheet on GPS-Galileo cooperation. Accessed November 9, 2014.
http://www.gps.gov/policy/cooperation/europe/2007/gps-galileo-fact-sheet.pdf.
U.S. Air Force. 2007b. 2007 Joint press release on civil signal design. Accessed November 9, 2014.
http://www.gps.gov/policy/cooperation/europe/2007/MBOC-agreement/.
U.S. Air Force. 2007c. 2007 Joint statement on GPS-Galileo working group B. Accessed November 9,
2014. http://www.gps.gov/policy/cooperation/europe/2007/working-group-b/.
U.S. Air Force. 2008. 2008 GPS-Galileo working group B meeting. Accessed November 9, 2014.
http://www.gps.gov/policy/cooperation/europe/2008/working-group-b/.
U.S. Air Force. 2010. 2010 Joint statement on combined performance of GPS and Galileo. Accessed
November 9, 2014. http://www.gps.gov/policy/cooperation/europe/​2008/joint-statement/.
158 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

U.S. Air Force. 2013. 2013 Interim report of ARAIM technical subgroup. Accessed November 9, 2014.
http://www.gps.gov/policy/cooperation/europe/2008/joint-statement/.
U.S. Air Force. 2014a. New civil signals. Accessed November 9, 2014. http://www.gps.gov/systems/
gps/modernization/civilsignals/#L1C.
U.S. Air Force. 2014b. Official U.S. government information about the global positioning sys-
tem (GPS) and related topics. Accessed November 9, 2014. http://www.gps.gov/policy/
cooperation/#europe.
Ward, P., J. Betz, and C. Hegarty. 2006. Interference, multipath, and scintillation. In Understanding GPS
Principles and Applications, 2nd edn., E. Kaplan and C. Hegarty (eds.), pp. 279–295. Norwood,
MA: Artech House.
Xie, P. and M. Petovello. 2015. Measuring GNSS multipath distributions in urban canyon environ-
ments. IEEE Transactions on Instrumentation and Measurement 64(2): 366–377.
7
Electromagnetic Interference to
Aeronautical Telecommunications

Zhigang Liu

CONTENTS
7.1 Introduction......................................................................................................................... 160
7.2 Concepts of Aeronautical Communication Electromagnetic Interference................. 160
7.2.1 Related Terms.......................................................................................................... 160
7.2.1.1 Electromagnetic Fields (EMFs)............................................................... 160
7.2.1.2 Electromagnetic Interference................................................................. 161
7.2.1.3 Electromagnetic Compatibility.............................................................. 161
7.2.1.4 High-Intensity Radiated Fields (HIRFs)............................................... 162
7.2.1.5 Lightning and Lightning Strike............................................................. 162
7.2.2 Civil Aviation Communication System Model................................................... 162
7.2.3 Aeronautical Communications EMI Classification and Basic Measures........ 164
7.2.4 Aeronautical Communications Laws and Regulations..................................... 165
7.2.4.1 Aeronautical Communications EMI History....................................... 165
7.2.4.2 Laws and Regulations in China............................................................. 166
7.3 Control of Aeronautical Communications EMI............................................................. 174
7.3.1 Effects of Aeronautical Communications EMI................................................... 174
7.3.1.1 EMI Impact on Aeronautical Communications................................... 174
7.3.2 Common Interference Analysis............................................................................ 176
7.3.3 Common Sources of Interference and Countermeasures................................. 177
7.3.3.1 Analysis of Interference Sources........................................................... 177
7.3.3.2 Remedy...................................................................................................... 179
7.4 Aeronautical Communications System Protection Requirements.............................. 180
7.4.1 Aircraft Design Protection Measures.................................................................. 180
7.4.1.1 Grounding................................................................................................. 180
7.4.1.2 Overlap...................................................................................................... 180
7.4.1.3 Shielding.................................................................................................... 180
7.4.1.4 Static Discharger...................................................................................... 181
7.4.1.5 Lightning Protection................................................................................ 181
7.4.2 Ground Station Design Protection Measures..................................................... 181
7.4.2.1 Airport Requirements............................................................................. 181
7.4.2.2 NDB Station Requirements..................................................................... 182
7.4.2.3 LOC Station Requirements..................................................................... 183
7.4.2.4 GS Station Requirements........................................................................ 184
7.4.2.5 MB Station Requirements....................................................................... 186
7.4.2.6 DVOR Station Requirements.................................................................. 187

159
160 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.4.2.7 DME Station Requirements.................................................................... 188

7.4.2.8 Precision Approach Radar Station Requirements............................... 188
7.4.3 Study Case............................................................................................................... 188
References...................................................................................................................................... 190

7.1 Introduction
This chapter mainly analyzes the electromagnetic interference (EMI) of aviation com-
munications equipment, introduces the EMI effects and response measures of aviation
communications equipment, and then analyzes the protective measures of aircraft and
ground stations against EMI; it also discusses EMI protection for aeronautical communica-
tions equipment.

7.2  Concepts of Aeronautical Communication Electromagnetic Interference

In the world, electromagnetic waves are everywhere. A variety of electronic devices such
as radio, broadcast television, communication transmitters, and other radar, navigation
equipment radiate electromagnetic energy into space during operation. This electromag-
netic energy is intentional radiation, because it is consciously made to transfer a v­ ariety
of data.
However, many electronic devices will also radiate electromagnetic energy into space
during operation, such as from switching circuits, automatic ignition systems, and auto-
matic control equipment. This electromagnetic energy is not deliberately manufactured,
as it automatically appears in the circuit work process. They may affect other electronic
devices, work and are undesirable, so is unintentional radiation.
Thus, because the electromagnetic environment (EME) is intentionally and unintention-
ally produced, it cannot be eliminated. When the intensity of electromagnetic radiation
rises to a certain extent, it affects the normal operation of other electronic equipment and
electronic systems. All aeronautical communication devices, in fact, must adopt appropri-
ate measures to protect against electromagnetic radiation.

7.2.1  Related Terms

7.2.1.1  Electromagnetic Fields (EMFs)
Electromagnetic fields are formed by electric and magnetic fields. Electric fields refer to
electric charges exerted on other electric charges. Magnetic fields refer to the magnetic
object or moving electric charge effect on other magnetic materials and electric charges.
EMFs take place throughout the electromagnetic spectrum; they are generated by natural
or human activity. Some EMFs occur naturally, such as Earth’s magnetic field, static elec-
tricity, and lightning. Also, others are established by the transmission and distribution
of electricity, used in household electric appliances, communication systems, industrial
processes, and scientific research.
Electromagnetic Interference to Aeronautical Telecommunications 161

The whole frequency spectrum is covered by natural and human-generated EMFs.

DC (direct current) EMFs are EMFs that are nearly constant in time, and AC (alternating
current) EMFs are those that vary in time. AC EMFs are characterized by their frequency
range. The lower limit of 3–3000  Hz is defined as extremely low-frequency magnetic
fields. Radio communications devices are operating at much higher frequencies, which
are often in the range of 500,000 Hz (500 kHz)–3 billion Hz (3 GHz). The radio frequency
(RF) sources of EMF typically include cellular telephone towers, broadcast towers for radio
and television, airport navigation, radar, and communication systems. The high-frequency
(HF) communication systems and very-high-frequency (VHF) communication systems
are used by emergency medical technicians, utilities, and governments. Examples of local
wireless systems are WiFi, cordless telephone, and so on.

7.2.1.2  Electromagnetic Interference

EMI is a repeated interruption or serious degradation of licensed radio communica-
tions service by any signal radiated in free space or conducted along signal leads, which
disrupts the function of radio navigation or other safety services. It occurs when the
EMFs produced affect operation of an electrical magnetic source adversely or other elec-
tromagnetic device. EMI can be caused by the source that intentionally radiates EMFs
or one that does so incidentally. EMFs can be described in terms of the frequency or the
times the EMFs changed direction in space each second. Radio communications services
are of a great variety, such as AM/FM commercial broadcast, television, telephone ser-
vices, navigation services, radar, and air traffic control (ATC). The licensed radio services
together with some unlicensed radio services, such as WLAN or Bluetooth, may inter-
rupt the EME.
Any EMI problem has three basic elements: interference source, path, and victim as in
Figure 7.1. This chapter focuses upon the EMI situation of the airplane communication
system. Because many interference sources are normally operated by battery power when
used on board airplanes, the path is primarily radiated, rather than conducted. The limi-
tations of such devices such as physical size and transmit power are given; the primary
concern of EMI is for aircraft communication systems (victim).

7.2.1.3  Electromagnetic Compatibility

Electromagnetic compatibility (EMC) is a concept wherein the equipment and systems
work properly in their EME and meet the requirements to run and at the same time do
not produce EMI in the electromagnetic disturbance. The definition contains two aspects:
First, the device should be able to work under certain normal EME and that the device
should have a certain electromagnetic immunity (EMS). Second, electromagnetic distur-
bance, which is generated by the device itself, should not have too much impact on other
electronic products, that is, EMI.

Interference Path
Victim
source

FIGURE 7.1
Three basic EMI elements.
162 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

In order to regulate the EMC of electronic products, all developed countries and some
developing countries have developed standards for EMC. EMC standards guarantee
that products can work in the actual EME of the basic requirements. It is called the basic
requirements, that is to say, even though the products meet the EMC standards, in actual
interference, problems may occur. Most national standards are based on the standards
developed by the International Electrotechnical Commission.

7.2.1.4  High-Intensity Radiated Fields (HIRFs)

HIRF is higher radiation energy per unit area of the electromagnetic radiation, which is
determined by the electric field and magnetic field strength of the electromagnetic wave.
HIRF is radiation from the ground, ships, offshore platforms, or aircraft on radar, and
radio, television, satellite uplink data, high-power transmitters. It is an EME problem
caused by human activity, which is characterized by frequency bandwidth and long dura-
tion of action. In aeronautical communication systems, many radio antennas are in HIRF
areas. The antenna is a radiator that converts the existing voltage–current form within
the transmitter to a free-space electromagnetic wave, so it can help accomplish aircraft
communications and navigation tasks. However, HIRFs are harmful to communication
systems and some accidents are caused by HIRF, which has become an important factor
affecting the safety of aircraft.1

7.2.1.5  Lightning and Lightning Strike

Lightning is a discharge of atmospheric phenomena, mostly formed in cumulonimbus.
Cumulonimbus with changes in temperature and airflow will continuously move, and
as the movement continues and friction powers, charged clouds are formed. Some clouds
are positive charged, while others are negative charged. In addition, buildings and trees
below the clouds accumulate charges due to electrostatic induction. With the accumula-
tion of electric charge, the voltage of a thundercloud gradually increases, and when the
thundercloud and the projections, ground with different electric charges, are adjacent to a
certain degree, that is, when the electric field exceeds 25–30 kV/cm, intense discharge will
occur, which is lightning, a simultaneously strong flash. Since the discharge temperature
is high, there is a rapid expansion of heated air followed by a roar of explosion, which is
thunder. In the lightning discharge process, objects that conduct lightning, that is, those
being struck by it, would be destroyed. Also, an aeronautical communications system is
affected by lightning transient EMI; hence, the design and operation of system should take
into account the impact of lightning.

7.2.2  Civil Aviation Communication System Model

Civil aviation communication services mainly include ground communications, mobile
communications in airport, telephone communications, ground–air communication, flight
operation management communication, on-air radio, and confidential communications.
The horizontal communications network is currently based on data communications net-
work such as X.25 FR public network, frame relay, and asynchronous transfer mode ATM.
Ground–air communication networks mainly include HF communications, VHF commu-
nications, ultra-HF (UHF) communications, and satellite communications.
The aeronautical communications system includes the VHF communication sys-
tems, HF communications systems, selective call system (SELCAL), intercom system,
Electromagnetic Interference to Aeronautical Telecommunications 163

automatic direction finder (ADF), VHF omnidirectional range (VOR), instrument land-
ing system (ILS), radio altimeter (RA), distance measuring equipment (DME), and ATC
transponder and weather radar (WXR). The VHF system uses 118–135.975  MHz bands,
channel spacing of 25 kHz, which is mainly for aircraft during takeoff and landing and
for aircraft in controlled airspace during a two-way voice communication between the
ground traffic control personnel. VHF signals use only linear wave propagation or line-of-
sight propagation. The HF system uses 2–30 MHz bands, channel spacing of 1 kHz, which
is mainly for aircraft communication with the airport during remote flight. The purpose
of remote communication via reflection is to extend the communication distance up to
thousands of kilometers between the ionosphere and the ground surface. SECAL is not
an independent communications system, which coordinates with VHF and HF systems.
The intercom system is used mainly for calls, broadcast, maintenance personnel calls, as
well as audiovisual signals to travelers. ADF uses the long wavelength of 100–2000 kHz,
which is for measuring ground navigation station with respect to the longitudinal axis of
the aircraft position and for guiding the aircraft to fly. VOR uses the frequency band of
108–118 MHz, channel spacing of 50 kHz, to measure the VOR azimuth that determines
the deviation of the aircraft relative to the selected VOR route. The ILS is the important
equipment to ensure the safety landing of aircraft, by the localizer (LOC), glide slope (GS),
and marker beacon (MB) components. The localizer uses 108.1–111.95 MHz bands, channel
spacing of 50 kHz, which guide the aircraft to be aligned with the runway during landing.
The glide slope uses 329.15–335.0 MHz bands, channel spacing of 150 kHz, which guide
the aircraft fall along 3° angle to the runway during landing. The marker beacon uses
75 MHz bands, which check the correct aircraft height and speed through the beacon and
provide the distance information of aircraft away from the runway threshold. RA uses
4200–4400 MHz bands, which are used to measure the actual height of the aircraft relative
to the ground surface at the time of the approach and landing. DME uses 1025–1150 MHz
and 962–1213 MHz bands, which are used to measure the distance of aircraft and DME
beacon. ATC transponder uses 1030 and 1090 MHz bands, which are used to report the
identification code and pressure altitude of aircraft to the ground control center and can be
used to determine the horizontal position of the aircraft. WXR uses 9330–9400 MHz bands,
which are used to detect typhoon target within the sector ahead of the aircraft and other
obstacles, then select a safe route around the obstacle.
The principle of civil aviation communications systems is to specify powers, frequency,
and antenna requirements, which ensure the reliable transmission of data across the wire-
less data link, while giving attention to the effects of propagation attenuation and noise
effects relating to the free-space transmission, as illustrated in Figure 7.2.
The main link parameters contain the link operating frequency, the output power of
transmitter, and the antenna parameters. In certain scenarios, increasing the output power
levels of transmitter can reduce certain aspects of wireless EMI, but the method may

Free-space transmission

Transmitter Receiver

FIGURE 7.2
Civil aviation communication link.
164 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

cause more unnecessary EMI to other electronic and communication systems, so it is not
practical for aviation communications devices with finite power sources.
The equations that can be used to calculate free-space transmission loss are given in the
following. The distance variables are directly proportional. When the value of distance
parameters is increased, the value of the free-space loss will increase at the same time.
The frequency variables are the same with distance parameters to the value of the free-
space loss:

L ( dB ) = 32.5 + 20log10 d + 10log10 f

where
L means the free-space transmission loss in dB
d means the free-space transmission distance in km
f means the frequency in MHz

There are many factors that affect the propagation of radio waves, such as transmission
power of transmitter, antenna gain, attenuation factors, and environmental noise. In addi-
tion, the ionospheric and tropospheric conditions of the environment also obviously take
effect on the propagation of radio transmission in free space.
The spread of HF signals depends on sky wave propagation when their distances
are approximately greater than 100 miles, depending on the ionization levels that exist
in the ionosphere. In most cases, civil aviation HF communications propagate along
the shortest path between the transmitter and receiver, and the signals propagated use
the lower atmospheric region. The effects of the ionosphere are more obvious to the
medium- and high-frequency regions signals than the VHF/UHF range signals, because
it is refracted over a considerable vertical distance gradually. The highest frequency that
is supported for reflection by the ionospheric layer is known as the maximum usable
frequency.
The spread of VHF/UHF signals depends on ground wave propagation, which is almost
always followed by line-of-sight paths. So it is required that there be no obstructions
between the transmitter and receiver. The limitations of distance in this frequency range
are primarily the functions of the antenna height of the transmitter and receiver.

7.2.3  Aeronautical Communications EMI Classification and Basic Measures

The International Telecommunication Union (ITU) is responsible for the allocation of fre-
quency spectrum, and the frequency spectrum is assigned by state administrations, which
are mostly government bodies. The assignments of each country are based on the alloca-
tions established in the ITU radio regulation. The assignment of frequency spectrum to
airport and other aeronautical communication systems is the responsibility of the state
civil aviation authority.
According to the allocation of ITU, aeronautical communication services are given
­primary status that is very important, so they must be protected from EMI from other com-
munication services. Because broadcasting services and mobile communication ­services
have secondary status, they must avoid causing EMI to any primary services and must be
shut down if necessary.
EMI of aeronautical communication systems is primarily caused by undesirable voltages
or current; they affect the performance of the aeronautical communication system and are
Electromagnetic Interference to Aeronautical Telecommunications 165

processed as noise. Different EMI sources can be identified and mainly can be categorized
into two kinds. One kind of EMI is from natural sources such as lightning, atmospheric
effects, electrostatic discharge, sunspot activity, and reflections from the rough Earth sur-
face. Other kinds of EMI are the man-made sources such as industrial activity, high tension
electric cables, radar, and broadcasting transmitters.
The victims receive electromagnetic energy from the source of interference, and this
energy has two main propagation mechanisms, radiation and conduction. EMI radiation is
further divided into direct radiation and third-order intermodulation products r­ adiation.
The direct radiation originates from communication equipment and its interference uses
the same frequency as the aeronautical communications system and radiates electromag-
netic energy. The third-order intermodulation products radiation occurs when system
operation creates one or more strong interference signals on certain frequencies and affects
aeronautical communication systems. Similarly, intermodulation interference may also be
generated in an aeronautical system receiver, which is caused by receiver being driven
into nonlinearity leading to high-power broadcasting signals outside the band. When the
interference is produced by two or more broadcasting signals that have a frequency rela-
tionship, an intermodulation product of the nonlinear process can occur within the RF
channel through the aeronautical receiver.
Many techniques and measures have been taken in order to reduce the effects of EMI in
communication systems. A variety of mitigation techniques have been used according to
the interference nature, interference source, and interference mechanism of different com-
munication services and systems. So not all mitigation technology measures are suitable
for all cases. ITU provided a recommendation for EMI protection of safety services from
undesirable emissions; many mitigation measures have been stipulated, which include the
following:

1. Improving RF selectivity in order to reduce undesirable signals outside of the

tuned bandwidth and provide good adjacent channel rejection performance.
Beam down tilt so as to reduce interfering signal and also increase penetration.
2. Modification of antenna pattern that is much facilitated by corner reflectors and
directional antennas, which when used direct a signal to the intended area of
interest, thus minimizing interference to outside the service area.
3. Deployment of electronic RF filters such as notch filter and band-pass filters.

7.2.4  Aeronautical Communications Laws and Regulations

7.2.4.1  Aeronautical Communications EMI History
It can be said that civil aviation and radio communication have developed and grown up
together. The development of radio communication technology has promoted the devel-
opment of civil aviation communication technology. Since the 1920s, radio communica-
tion technologies have continually adapted to the ever-expanding need for aeronautical
navigation and communication capability. At the same time, the need for reliable, secure
communication, navigation, and surveillance (CNS) and air traffic management (ATM)
for aeronautical communication has continually driven improvements in radio technol-
ogy. Because civil aviation has become more dependent on radio communication for CNS,
the EMI effects have become ever more critical and urgent. With very few exceptions, the
aeronautical radio spectrum has been affected by other radio services, which have been
coordinated and protected by law to minimize the potential of EMI. However, because
166 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

EMI have resulted in numerous aviation incidents and accidents over the years, which are
effects of lightning, electrostatic discharge, and HIRFs from radars and broadcast transmit-
ters, so EMI is carefully considered in all aspects of design now.
Since the inception of aviation, lightning has always been a problem, and flight mis-
sions need to occur in all weather conditions now. The National Advisory Committee for
Aeronautics (NACA) established the Aircraft Safety Subcommittee in 1938. Lightning
effects on aircraft have been studied by weather and lightning experts to determine what
additional protective measures are needed. There is an influential history of airplanes
struck by lightning, sometimes with catastrophic effects. Today, lightning strikes have
minimal impact on the safety of airplanes, which has been well established by the FAA
lightning protection regulations. The thorough technical guide to protecting aircraft
from the effects of lightning has been provided by the textbook Lightning Protection of
Aircraft.
In the early 1960s, compact radio receivers became available, which featured new
­transistor circuitry and were carried on airplanes by passengers. These devices could dis-
rupt the ILS, VOR, and other navigation systems. In order to solve this concern, Special
Committee 88 was formed by the Radio Technical Commission for Aeronautics (RTCA),
which published Interference to Aircraft Electronic Equipment from Devices Carried Aboard
(RTCA/DO-119) in February 1963. This began to control the use of compact radio receivers
on board aircraft; this report recommended prohibiting the operation of portable radios
during flight and set the stage for restrictions on passenger-carried compact radio receiv-
ers that are still in effect today.
By the 1970s, the problem of HIRF emanating from broadcast towers, radars, and point-
to-point radio links arose; the new digital flight control systems needed to be robust against
the effects of HIRF. The Department of Defense began to take into account for the control
of EMI on aircraft systems in the United States. Many EMI-related handbooks relating
to RF environment definition, procurement guidelines, hazards, grounding, bonding, and
shielding exist. For commercial civil aviation, the RTCA worked with the FAA to set envi-
ronmental standards and test procedures for aircraft equipment. Since the 1980s, NASA
has documented several HIRF- and EMI-related events studies.

7.2.4.2  Laws and Regulations in China

With the development of China civil aviation, EMI to communication systems has become
an urgent problem. The State Council and the Central Military Commission issued pro-
visions on the protection of airport clearance in 1982, which specify the requirements of
field emergency takeoff runway, airport, and clearance area between the airport runway
and the highway. Then Electromagnetic Environment Requirements for Aeronautical Radio
Navigation Stations (GB6364-1986) was enacted in May 1986 as the national standard in
China.2 This document analyzed the effect from nonaeronautical navigation services on
all types of radio equipment, which include high-voltage power lines; electric railways;
and industrial, scientific, and medical equipment, such as jamming caused by coupling
the navigation station reflection or reradiation of the surrounding terrain; these factors
may have harmful impacts on navigation information. This document lists the radio site
protection requirements and measurement methods of many communication and navi-
gation systems, which include nondirectional beacon (NDB), VHF/UHF DF, ILS, VOR,
DME, TACAN, and PAR stations.
Electromagnetic Interference to Aeronautical Telecommunications 167

In August 1996, the Civil Aviation Administration of China (CAAC) released “Aeronautical
radio navigation aids and ATC radar station sitting criteria” (MH/T 4003-1996). The indus-
try standard standardized navigation, radar, and ATC radar terminal system setup loca-
tion site environmental requirements, which is according to the civil aviation law, the
GB6364-1986 national standard, and the International Civil Aviation Organization Annex
10 and other related articles of the convention on international civil aviation. It is related
to terrain and houses, trees, embankments, and other active and passive objects, active
and passive jamming, and the dual nature of overhead power lines, electric railways, and
so on. In this document, the protections for radio interference are aimed at the various
communications and navigation systems. It stipulates more detailed and specific require-
ments of ATC short-range and remote-range surveillance radar and ATC secondary radar
stations. In 1997, the CAAC radio management committee gave notification of designated
civil aviation airport EME protection areas, which reduce harmful interference to civil
aviation radio stations.
In August 2010, “Electromagnetic environment protection requirements and mea-
surement methods for VHF/UHF band radio monitoring station” (GB/T 25003-2010)
was enacted as a national standard in China, which provides VHF/UHF band electro-
magnetic disturbance environment that allows values for radio stations and surround-
ing obstacles limiting requirements, provides for test method of EMI field strength,
and defines the VHF/UHF band radio stations surrounding spacing requirements for
minimum protection of high-power sources. The CAAC released “VHF air–ground
communications ground system” (MH/T 4001.1-2006), which describes the VHF EME
requirements. Likewise, “HF ground-to-air communications equipment general specifi-
cations” (MH/T 4002.1-1995) and “Electromagnetic environment requirements for short-
wave radio receiving stations” (GB13617-92) were enacted, which describe the HF EME
required of the receiving station. In December 2011, the ATC industry office of the CAAC
released “Specification for civil aviation airport electromagnetic environment protec-
tion area delineation and protection requirements” (AC-118-TM-2011-01), which provides
EME protection area delineation and protection requirements for civil aviation airport.
In April 2013, the ATC industry office of the CAAC released “Civil airports and air traffic
radio electromagnetic environment test specification” (AP-118-TM-2013-01), which stan-
dardized test civil airport and EME for air traffic radio stations.
With the rapid development of the civil aviation industry, the original GB6364-86 civil
specifications have been unable to meet aviation EME protection requirements; the con-
tradictions of urban construction and proliferation of navigation stations are increas-
ingly prominent. Most of the standards currently cannot meet the development of civil
aviation technology; the execution results are unsatisfactory. The rapid development of
high-voltage power lines, power plants, roads, railways, radio communications has only
increased radio interference. Also, in airport construction, EME issues have become
increasingly prominent. The appropriate revision, “Electromagnetic environment require-
ments for aeronautical radio navigation stations” (GB6364-2013) was enacted in December
2013 as a national standard in China, which takes into account the development of mili-
tary and civil aviation and EME requirements, taking into account the GB line with ICAO
standards compliance. The practicality and operability of new standards is stronger and
is conducive to EME.
HF/VHF EME protection requirements include VHF EME requirement and HF receiv-
ing station EME requirement.
168 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.2.4.2.1  VHF EME Requirement

The setting requirements for VHF aviation ground station EME are given in the
­following table:

Protect the signal electric field intensity (dBμV/m) 26

Electric field intensity Protective rate (dB) FM broadcasting 17
Other 15
FM broadcasting Open value −10
Interference threshold (dBmW) Cutoff value −30

Protection distances of all kinds of interference sources for civil aviation VHF ground
­station are given in the following table:

Interference Sources Protection Distance (m)

FM broadcasting 1 kW (include) less than 1000
1 kW more than 6000
Electrified (electric) railways 300
Arterial highway Level 2 and above highway 300
High-voltage transmission line 110–220 kV 300
220–330 kV 250
500 kV 300
Industry–science–medicinal Permissible variation meets the 800
RF equipment requirement

7.2.4.2.2  HF Receiving Station EME Requirement

HF receiving station EME must meet the MH 4002.1 and GB 13617-92 standards and EME
requirements for shortwave radio receiving stations as given in the following table:

Protection Distance of Different Levels of Stations (km)

Interference Sources First Level Second Level Third Level
Middle- and long-wave <50 10 7 3
transmitting station (kW) 100–150 15 10 5
>200 12 12 7
High-voltage transmission 500 1.8 1.1 0.7
line (kV) 200–330 1.3 0.8 0.6
110 1.0 0.6 0.5
Shortwave transmitting 0.5–5 4 2 1.5
station (directions in 5–25 4–10 2–6 1.5–3
communications ¼ power 25–125 10–20 6–10 3–5
angle) (kW)
>120 >20 >10 >5
Shortwave transmitting 0.5–5 2 1 0.7
station (directions in 5–25 2–5 1–3.5 0.77–1.5
noncommunications ¼ 25–125 5–10 3.5–5 1.5–2.5
power angle) kW
>120 >10 >5 >2.5
Motor road High-speed, 1 0.7 0.5
first level
Second level 0.8 0.5 0.3
Industry–science–medicinal Common 2 1.4 0.7
RF equipment More powerful 5 3.5 0.5
Electromagnetic Interference to Aeronautical Telecommunications 169

First-level short radio transmitting stations include the following:

1. Communication of central authorities, the state council, committee, frame, and

the Chinese People’s Liberation Army units that monitor operating department
transmitting stations.
2. For defense, public security, and national safety and maritime security at stake
and large building facilities or underground fortifications and trenches, massive
antenna sites, and large high-gain antenna receiving stations.
3. Communication objects other than Asian countries, regional or distant-water
fleets of receiving stations.
4. If a receiving station is located in the coherent region or urban areas, it must accord
to the secondary-level station protection even though it is compliance with the
aforementioned condition.

Second level shortwave radio receiving stations include the following:

1. Provinces, autonomous regions, and municipalities directly under the central

receiving station, and People’s Liberation Army–level unit directly under the
receiving station
2. Asian countries such as Taiwan and its maritime communications.
3. The stations that meet these conditions include aeronautical communications
and radar stations, navigation stations, and shortwave radio receiving stations.
However, all three stations have equal amplitude reception.

Third-level shortwave radio receiving stations include receiving stations that were set by
county or province, which include formal erection of antennas and other construction and
which are responsible for more important communication tasks.3
Protection of the long-wave transmitter station spacing:

Protection Spacing (km)

Transmit Power (kW) First-Level Station Second-Level Station Third-Level Station
<100 10 7 3
100–200 15 10 5
>200 20 12 7

Protection of the shortwave transmitter station spacing:

Protection Spacing (km)

Transmit Power (kW) First-Level Station Second-Level Station Third-Level Station
0.5–5 4 2 1.5
5–25 4–10 2–6 1.5–3.0
25–120 10–20 6–10 3.0–5.0
>120 >20 >10 >5.0
170 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Protection beyond the ¼ power angle of the directional antenna power:

Protection Spacing (km)

Transmit Power (kW) First-Level Station Second-Level Station Third-Level Station
0.5–5 2 1.0 0.7
5–25 2–5 1.0–3.5 0.7–1.5
25–120 5–10 3.5–5.0 1.5–2.5
>120 >10 >5.0 >2.5

Shortwave radio receiving station (station) on the protection of high-voltage overhead

transmission wire spacing:

Protection Spacing (km)

Voltage Level (kV) First-Level Station Second-Level Station Third-Level Station
500 1.8 1.1 0.7
220–330 1.3 0.8 0.6
110 1.0 0.6 0.5

Road protection spacing:

Protection Spacing (km)

Road Level First-Level Station Second-Level Station Third-Level Station
High-speed, first level 1.0 0.7 0.5
Second level 0.8 0.5 0.3

Protection pitch for industrial, scientific, and medical RF device users borders:

Protection Spacing (km)

Industry, Science,
Medicinal RF Device First-Level Station Second-Level Station Third-Level Station
Common 2.0 1.4 0.7
More powerful 5.0 3.5 1.5

Aeronautical radio frequency bands corresponding of radio stations:

Station Category Radio Frequency Bands Polarization

Communication HF 2.8–22 MHz Vertical
VHF 118–137 MHz Vertical
Satellite Earth C-band 3,968–3,991 MHz (Uplink) —
Station Ku-band 12,688–12,742 MHz —
(Downlink)
(Continued)
Electromagnetic Interference to Aeronautical Telecommunications 171

Station Category Radio Frequency Bands Polarization

Navigation NDB 190–900 kHz Vertical
ILS LOC 108–112 MHz Horizontal
GS 328.6–335.4 MHz Horizontal
MB 75 MHz Horizontal
DVOR 108–112 MHz Horizontal
DME 960–1215 MHz Vertical
Surveillance Primary radar Remote 1,250–1,350 MHz Linear
Process 2,700–2,900 MHz
Secondary radar 1,029–1,031 MHz Vertical
1,087–1,093 MHz
ADS-B 1,089–1,091 MHz Vertical
Meteorology Boundary layer wind 1,270–1,295 MHz —
profiler radars 1,300–1,375 MHz
Weather S-band 2,700–2,900 MHz —
radar C-band 5,300–5,600 MHz —
X-band 9,300–9,700 MHz —

Airport maximum permissible interference field strength and protection spacing

requirements:

Maximum Permissible Interference

Station Category Field Strength
Communication HF 20 dBμV/m
VHF 9 dBμV/m
Navigation NDB 22 dBμV/m (outside latitude 40°)
27 dBμV/m (in latitude 40°)
ILS LOC 41 dBμV/m
GS 12 dBμV/m
MB 32 dBμV/m
DVOR 19 dBμV/m
DME 55 dBμV/m
Surveillance Primary radar −139 dBm
Secondary radar −106 dBm
ADS-B −85 dBm
Meteorology Boundary layer wind profiler −175 dBm
radars
Doppler weather radar S-band −151 dBm
C-band −150 dBm
X-band −150 dBm
Common weather radar C-band −150 dBm
X-band −145 dBm
Airport terminal S-band −153 dBm
area, Doppler weather C-band −159 dBm
radar X-band −159 dBm
172 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Protection spacing requirements of communication equipment:

Maximum Permissible
Interference Field Strength Protection Spacing Requirements
Station Sources of Allowed Values Sources of Protection
Category Interference (dBμV/m) Interference Category Pitch (km)
HF — 20 Medium-wave and <50 kW 10
long-wave transmitter 100–150 kW 15
station >200 kW 20
Shortwave transmitter 0.5–5 kW 4
station 5–25 kW 4–10
(communication 25–120 kW 10–20
direction within 1/4 >120 kW >20
power angle)
Shortwave transmitter 0.5–5 kW 2
station 5–25 kW 2–5
(communication 25–120 kW 5–10
direction outer 1/4 >120 kW >10
power angle)
High-voltage 500 kV 1.8
transmission line 220–330 kV 1.3
110 kV 1.0
Automobile road Highway 1.0
Slow road 0.8
Electrified railway 0.8
Industrial, scientific, Common 2.0
and medical radio Multiple power 5.0
frequency equipment
VHF FM 9 FM 1 kW and less 1.0
From 1 kW 6.0
Electrified railway 0.3
Automobile road 0.3
Others 11 High-voltage 500 kV 0.3
transmission line 220–330 kV 0.25
110 kV 0.2
Industrial, scientific, and medical radio 0.8
frequency equipment

Protection spacing requirements of navigation equipment:

Maximum Permissible Interference Field Strength (dBμV/m)

Industrial, Scientific, and Medical Radio Frequency
Station Category FM Equipment Others
NDB Outside latitude 40° — 28 22
In latitude 40° — 33 27
ILS LOC 15 18 12
GS — 38 32
MB — — 41
DVOR 22 25 19
DME — — 55
Electromagnetic Interference to Aeronautical Telecommunications 173

Protection spacing requirements of surveillance equipment:

Maximum Permissible
Interference Protection Pitch
Station Sources of Allowed Sources of Protection
Category Interference Values (dBm) Interference Category Pitch (km) Remarks
Primary High-voltage −139 High-voltage 500 kV 1.0
radar overhead overhead 220–330 kV 0.8
transmission transmission 110 kV 0.7
lines lines
Transformer −139 High-voltage 500 kV 1.2
substation transformer 220–330 kV 0.8
substation 110 kV 0.7
Electrified −136 Electrified railway 0.7
railway Nonelectrified railway 0.5
Automobile −136 Automobile road 0.7
road
HF heat −136 HF heat sealing machine 1.2 Counting
sealing from the
machine plant
HF furnace −134 HF furnace ≤100 kW 0.5 Counting
from the
plant
Industrial −134 Industrial ≤10 kW 0.5
welding welding
HF therapy −134 HF therapy ≤1 kW 1.0 Counting
machine machine from the
workplace
Farm power ≤1 kW 0.5
equipment
Secondary — −106 Same with the primary radar
radar
ADS-B — −85 Same with the primary radar

Protection spacing requirements of meteorology equipment:

Station Category Maximum Permissible Interference (dBm)

Doppler weather radar X-band −150
C-band −150
S-band −151
Common weather radar X-band −145
C-band −150
Airport terminal area, X-band −159
Doppler weather radar C-band −159
S-band −153
Boundary layer wind profiler radars −175
174 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.3  Control of Aeronautical Communications EMI

7.3.1  Effects of Aeronautical Communications EMI
7.3.1.1  EMI Impact on Aeronautical Communications
Each airport has its own set of primary and standby frequencies for communication in
the VHF AM, aircraft, and from various airports, while aircraft must frequently change
frequency in deferent airport stations, so that the bandwidth of the antenna must be a full
band. Aircraft design must limit the shape and number of antennas so as to reduce drag
and conform with the principles of aerodynamics. A full-band antenna can be used as
much as possible; generally this reduces antenna performance related to signal attenuation
of outside frequencies when compared with the single-frequency antenna. In the practical
application, VHF radio communication uses the same vertically polarized wave, which
cannot reduce co-channel interference when compared with directional antenna VOR/ILS
communication systems. An AM receiver can choose FM frequency signal with a linear
transformation as with processing for AM signal, and when frequency band is suitable, the
AM signal will be demodulated to sufficient strength.
VOR and ILS receivers use horizontal dipole antenna and horizontal polarization wave,
so there is a strong directionality. Polarization properties of the antenna determine cer-
tain attenuation to ground-based communication signals with vertical polarization. The
low-frequency demodulation processing circuit of VOR and ILS receivers processed 30,
90, or 150 Hz fixed frequency signals, so it is easy to filter out conventional 300–3000 Hz
voice signal. The marker beacon receiver used a fixed transducer capacitance to match
the shortened horizontally polarized antenna. This measure does not only have a vertical
polarization signal damping effect to the communication systems, but also the sensitivity
of the receiver is lower and the gain of the interference signal is smaller. This unique work-
ing way of VOR and ILS navigation system determines that they have strong antijamming
capability, thus avoiding interference.
EMI directly affects aeronautical communication and navigation systems; the conclu-
sion can be drawn by observing anomalous aircraft system behaviors that may result from
EMI. A general description of EMI effects for several aeronautical communication and
navigation systems observed during the EMI signal is provided there, and operational
methods to handle these effects are explored.
The effects on VHF communication system are varied according to specific modulation
types of EMI signal, as well as the particular aircraft radio model. Some VHF systems
were silent without any alerting to the crew or any interference indication prior to reach-
ing the threshold suddenly. Other VHF systems were subject to distortion and undesirable
noise when the power level of the interfering signal was increased, and they were judged
to be unusable by the pilot for voice communication. Without specific failure indications
or warnings during EMI testing, signal occurred for the VHF system. Failure of the VHF
­system occurred when the EMI signal was transmitted from numerous locations in the
passenger cabin, and their manner and level was similar. Thus, if EMI is occurring in situ-
ations of undesirable noise and selection of an alternate VHF communication system, it
would not likely mitigate the effect of EMI.
For ILS (localizer and glide slope) and VOR, the EMI signal can cause visible variations
in both vertical and horizontal indicators. There are two types of variations: offsets and
fluctuations of the course and glide slope. The navigation display has the indicator of
declaring operational EMI failure. But if this indicator has been indicated by a failure flag,
Electromagnetic Interference to Aeronautical Telecommunications 175

it may show that either the reference signal was too strong or the received signal was too
weak. VOR has an indicator of system EMI failure, thus the localizer or glide slope failure
flag appears on the navigation display.
The most common effect of EMI on DME was on displayed DME data. These effects were
generally more serious on one particular aircraft system, therefore, operational procedures
may be helpful for mitigating EMI impact. On some using live DME signals systems, the
indicated distance to that site was varied up to 1 nautical mile when an EMI signal was
present. The Morse code identifiers were affected sometimes. Therefore, the Morse code
identifiers cannot be verified if the indicated direction was disturbed.
For the ATC transponder, operational EMI failure is defined as a drop below 90% in
aircraft replies to interrogations. It may be helpful for operational procedures to include
checks of the ATC reply indicator if EMI is suspected. EMI effects to ATC receivers
gradually increased with higher EMI signal power levels. Thus, EMI is more likely
to affect system processing of ATC replies from distant airplanes rather than nearby
airplanes.4
The allowed values are shown in the following table, which are aeronautical communi-
cations systems for industrial, scientific, and medical equipment interference and attenu-
ation characteristics.

Frequency Protection Interference Interference Allowed

Protection Target Range (MHz) Rate (dB) Attenuation Rate Values (dBμV/m)
NDB 0.15–1.75 9 d−2.8 85
LOC/GS/VOR 108–400 14 d−1 40

Using the following formula, we can calculate the interference protection zone for
­industrial, scientific, and medical equipment:
⎛ E30 −ES +R ⎞
⎜⎝ 20A ⎟⎠
d = 30×10

where
d means the protection distance, that is, industrial, scientific, medical equipment from
the ground or airborne receiving equipment
E30 means interference allowed values of industrial, scientific, and medical equipment
ES means signal strength of protection target
R means protection rate
A means protection against interference in a decaying exponential decay rate

The signal strength of radio equipment is calculated as follows:

⎛d ⎞
EX = ES + 20lg ⎜ s ⎟
⎝ dX ⎠

where
EX means signal propagation field strength
ES means the known signal strength
dX means the propagation distance
ds means the distance from the signal source
176 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.3.2  Common Interference Analysis

The basic types of aeronautical communications and navigation frequency interference
are intermodulation interference, right-of-band interference, and co-channel interference.
Intermodulation interference is divided into the transmitter intermodulation interfer-
ence and receiver intermodulation interference.
Transmitter intermodulation interference occurs when the signals of multiple transmit-
ters fall in the bands of another transmitter. In the final amplifier nonlinear interaction,
unnecessary combined frequency of the received signal appears, which causes interfer-
ence to the received signal with the same frequency as the combination.
The transmitter intermodulation interference can be reduced by increasing the transmit-
ter coupling loss (antenna isolation). We can take the following measures:

1. Increase the distance between the transmitter antennas. In order to meet the iso-
lation >50 dB, the transmitter antenna should be placed vertically, spaced apart
about 3λ, or horizontally, spaced apart about 40λ.
2. Unidirectional isolator. Inserting high-Q band-pass filter between the transmitter
output and feeder, increasing the frequency of isolation.
3. Try to avoid intermodulation frequency group assignments in tall buildings,
mountains, and tower positions.

The intermodulation interference of the receiver refers to intermodulation frequencies

along with the single or multiple strong signals simultaneously entering the receiver. The
interference occurs in the front-end nonlinear circuits of the receiver (e.g., HF amplifier,
mixer). If the intermodulation frequencies fall into the receiver intermediate frequency
band, interference occurs.
The most common approach is to add a high-Q value of the resonator filter and improve
the selectivity of the front of input circuit, to suppress the signal outside the receiver
frequency.
Right-of-band interference is generated by the stray radiation of transmitter and spuri-
ous response of receiver interference. At low frequencies, parts of VHF and UHF mobile
communications equipment use the crystal oscillator to obtain a higher-frequency sta-
bility, particularly the base station transmitter. The desired transmit frequency ft can be
obtained by repeated octave of main vibration frequency f0. Because of the nonlinear and
multiplier amplifier effect, this results in a large number of harmonic frequencies. If the
frequency characteristic of filter circuit is poor, these harmonics will be amplified and
radiated together with ft, to interference operation of the receiver in the corresponding
frequency. The interference can be generated on one or more frequencies near the outside
of the transmitter bandwidth. In short, the interference is due to the value of the transmit-
ter spurious radiation caused. So all types of transmitter spurious radiation values have
stringent requirements in national standards. According to the provisions of the National
Transmitter Technical Specifications: when the qualified transmitter carrier power is
greater than 25 W, the radiated power of discrete frequencies should be less 70 dB than
the transmitter carrier power. If the transmitter spurious radiation is too high, it usually
introduces many problems, such as too much multioctave frequency, poor selection of
multiplier output circuit, poor shield isolation between the multiplier and other factors,
and so on.
In addition to the useful signals that were received by the receiver, other unwanted sig-
nals were received at other frequencies. The kind of response capabilities of other unwanted
Electromagnetic Interference to Aeronautical Telecommunications 177

signals, commonly referred to as spurious response, is related with the frequency purity of
the receiver local oscillator. Spurious response of the superheterodyne receiver was mainly
IF and mirror-frequency response. The IF response is caused when the interfering signal
frequency is equal to the intermediate frequency of the receiver, the interference signal
from the receiver input circuit, then leaks into the HF amplifier circuit; thus, its suppression
is not enough, so IF signals go directly to the circuit. Mirror-frequency response is caused
by subtracting local oscillator frequency from the image frequency, through the IF loop.
Therefore, the receiver can respond in the image frequency. Mirror-frequency interference
has occurred in military and civil aviation communications and navigation frequencies.
Because aviation communications and navigation radio receiver selectivity standards are
poor (spurious response restrain, adjacent channel selectivity) and cause mirror interfer-
ence easily, installation of narrowband filter at the input of the receiver is the common way
to handle mirror-frequency interference.
Co-channel interference occurs when its frequency is the same or close to the interfer-
ence frequency and the desired signal frequency and when the signal is received in a
similar manner.

7.3.3  Common Sources of Interference and Countermeasures

7.3.3.1  Analysis of Interference Sources
With the application of a wide variety of radio and nonradio equipment in recent years, the
interference of aviation communications and navigation equipment has increased.

7.3.3.1.1  Radio Equipment

1. High-power cordless phones: High-power cordless phones can interfere with avia-
tion frequencies, through co-channel interference. First, their transmit power is
too large, resulting in spurious radiation falling into the frequency spectrum
dedicated to aeronautical communication. Second, direct use of such telephones
will result in emission of frequency dedicated to aeronautical communications
and navigation. Illegal setup and use of high-power cordless telephones poses
greater risk, because it will cause random interference effects.
2. FM radio/TV: First, the 88–108  MHz band of FM radio is adjacent to the dedi-
cated 108–137  MHz band of aeronautical communication and navigation sys-
tem. Second, FM radio transmitter power is too stong, and numerous broadcast
antennas are located on the same mountain and towers, so it is easy to generate
intermodulation signals. Third, the cable additions channel uses the dedicated
frequency of aviation communications and navigation system directly. There are
multiple cable links, trunk amplifiers, cable amplifiers, and splitters, in TV lines. A
bad cable connection, damaged shielding layer, or cable signal leakage that reaches
a certain intensity will cause harmful interference to the dedicated frequency of
aeronautical communications and navigation.
3. Illegal radio transmitter: First, the power of an illegal radio transmitter is too strong.
Second, its frequency assignment is illegal, and its station set is irrational. Signals
of different frequencies are transmitted simultaneously to form intermodulation
products that fall within the aeronautical communications band. Third, an aging
transmitter equipment usually fails, and decreased work performance indicators,
mainly spurious emission indicators, fail, causing the right-of-band radiation to
affect aircraft communications.
178 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

4.
Mobile phone base stations and repeaters: A frequency of 800 MHz or more is gener-
ally close to the harmonic frequencies of airport navigation devices, so intermodu-
lation and other spurious signals possibly cause interference to airport navigation
devices.
5.
Aviation devices: Many radio components are used in aviation, and they are prone
to interference. There are main causes of interference: first, equipment failure or
spurious emissions and second, equipment installation configuration is not stan-
dardized, resulting in emission power shortage, insufficient sensitivity, or mul-
tiple devices affecting each other.

In addition, with the rapid development of wireless technology and the increasingly
wide range of wireless applications, some new sources of radio interference will also
appear.

7.3.3.1.2  Nonradio Equipment

Nonradio equipment, found in industry, science, and medicine, will emit a radio signal
during operation, and the RF signal is generally wide, and its signal strength large. But
these signals are not fed through amplification systems, and therefore, the attenuation
is obvious to disappear after the propagation distance of several hundred meters. In
theory, they do not often interfere with the aircraft, but if they are at a short distance
facing empty stations, they will interfere with aviation communications and navigation
frequencies.
As the high-voltage power transmission line voltage level continues to increase, the wire
surface corona discharge and other opportunities also increase, and the effect of corona
discharge produces radio interference. The essence of radio interference is corona dis-
charge process, and it produces some harmful electromagnetic waves of relatively wide
frequency from the low 50 Hz to an HF range. These frequencies can interfere with normal
operation of surrounding radio communication devices.
Radio interference of the transmission line is mainly caused by the wire corona dis-
charge; the surface contamination of the insulator also causes spark gap, clamp corona,
spark discharge, and glitches discharge. The radio interference of an electrical substation
is caused by many factors and is complex. It will have a major spark discharge switch
operation from the substation at normal operating and fault conditions. It also causes par-
tial discharge due to damage, contamination, and other reasons such as corona discharge
of bus and device, etc.
Transmission line radio interference includes uniform interference, nonuniform inter-
ference, and pulse interference; in theory, any radio device will be affected by frequency
interference, in fact, mostly, AM radio, communications (0.5–12  MHz), and television
interference.
The degree of interference to radio transmission mainly depends on the distance from
the transmitter device, performance of the communication device, the antenna orientation,
and transmission lines of various parameters, such as wire arrangement, level voltage
tower height, size sag, and weather conditions, such as sun, rain, or fog.
AC overhead transmission lines cause radio interference limits under normal circum-
stances, 110–500 kV high-voltage lines can generate a frequency of 0.15–30 MHz of radio
interference. AC high-voltage overhead transmission lines characteristic of radio interfer-
ence distance can be represented by the formula
Electromagnetic Interference to Aeronautical Telecommunications 179

400 + (H − h)2
EX = E + k ⋅ log
X 2 + (H − h)2

where
EX means interference field strength
E means interference field strength, 20 m from the line
X means the distance from the line
H means the line height from ground
h means the measuring instrument height from ground
k means attenuation coefficient

7.3.3.2 Remedy
Because ionospheric interference must be reflected back to the ground, thus, the height and
the electron density of the ionosphere changes with the seasons, day and night. Shortwave
communications will be affected by distance and signal strength. In order to ensure stable
and reliable shortwave communications, the appropriate frequency should be selected
based on factors such as seasons, day and night, and area. Generally, a higher frequency
should be selected in summer and daytime for appropriate communication; a lower fre-
quency should be suitable for winter and night.
To reduce the effect of the dead zone, one may increase the power of the ground sta-
tion to increase the wave propagation distance. The operating frequency can be reduced
to shorten the distance between the nearest sky wave reflections, thereby narrowing the
scope of the dead zone, and also improve flight altitude, because the higher the altitude,
the smaller the dead zone range.
To increase the communication distance, the VHF antenna should be mounted higher.
In air–ground communications, as aircraft fly higher, that is, the distance between the
aircraft and the ground is greater, the farther the ultrashortwave propagation distance is.
Safe professional use of radio frequencies is the premise and foundation to ensure
flight safety, so it must attach importance to radio management. In particular, airport
communications departments should increase the protection of the EME around the air-
port, rely on local government and local radio management departments, and formulate
relevant regulations through the legislative branch of government. Airport communi-
cations departments should strengthen communication and cooperation with the local
radio management departments. If radio interference is found, one should first analyze
the intensity, orientation, and type of the interference source and then seriously make
a record and provide a basis for finding the source of interference. Second, one should
appeal to the local radio management departments in a timely manner. One should
actively seek the establishment of monitoring equipment at the airport by local radio
regulatory authorities, so that aeronautical frequencies can be monitored to detect inter-
fering signals. Industrial and commercial administrative departments should intensify
efforts to investigate and ban the production and sale of illegal radio equipment.
Aeronautical radio equipment, being mobile, causes mutual interference unavoidably.
The interference may be caused by improper operation of personnel or equipment, so
authorities must take scientific analysis seriously, strengthen management in strict accor-
dance with equipment operation and maintenance procedures, conduct a comprehensive
180 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

inventory of radio frequencies and equipment, punish illegal use of radio equipment, and
subject equipment for testing on a regular basis.5

7.4  Aeronautical Communications System Protection Requirements

7.4.1  Aircraft Design Protection Measures
As the complexity of electronic equipment and electrical systems increase, interference of
airborne electronic equipment is becoming increasingly serious. Therefore, it has become
a very important task to eliminate such interference, so that all airborne electronic equip-
ment can be working in conditions of EMC. The main measures taken are as follows:

7.4.1.1 Grounding
For radio equipment, the aircraft fuselage structure serves as the ground, which is the
common port. In order for the radio device to perform its function properly, it should
maintain an appropriate balance between aircraft and antenna structure. That is, the sur-
face potential of the aircraft should be stable. However, the operation surface of the aircraft
is separated from the rest parts sometimes, so it causes difference by body surface and
operation surface potential. If we do not alleviate the situation by grounding, it will affect
the operation of radio equipment. Therefore, it is connected between the plane operating
surface and the ground.

7.4.1.2 Overlap
An aircraft in flight will generate a lot of static friction between airframe and dust in the
air, water, and other particles in flight. If there is poor contact between the parts of the
airplane, the charge generated by each metal member will vary, so that there is a cer-
tain potential difference between the plane metal surface. When the potential difference
reaches a certain value, it will cause an electrostatic arc discharge. Arcing occurs in the
form of short pulses, which generate noise and will cause EMI in the radio spectrum to
antennas and other radio equipment.
Overlap can provide a low-resistance path for all metal parts within the aircraft,
­eliminating the potential difference between the metal parts, thereby eliminating EMI
caused by electrostatic discharge, but also maintain a constant flow of low-resistance path
on the charge. It is used to connect devices that have lap belts and lap folders commonly.
Overlap between the radio and the airplane structure provides a low-resistance path
between the common port, which also reduces static interference.

7.4.1.3 Shielding
Shielding is one of the most effective methods of removing noise; its basic purpose is to
shield the RF noise in electrical energy, which reduces the electromagnetic energy radiated
outward. For example, the use of shielded antenna feeder cable can reduce the radiation
of electromagnetic energy outside and can prevent external electromagnetic energy from
entering into the feed line. The data bus on the aircraft also require mask processing and
proper grounding.
Electromagnetic Interference to Aeronautical Telecommunications 181

In a radio device, those circuits that radiate HF electromagnetic energy are used together
with the shield cover to prevent them from affecting other circuits. Electromagnetic radia-
tion energy from the inner surface of the shield is conducted into the ground. In the case
where the filter cannot be used, shielding is particularly effective. For example, when the
electromagnetic radiation source radiates energy, the receiver input circuit and other vari-
ous circuits that are connected will receive the noise. In this case, we should install filters
on each affected line, which is virtually impossible, so it is best to take radiation shield-
ing into effect, because it will bar the radiation source that is enclosed within the shield,
thereby reducing the electromagnetic energy radiated outward.
For example, ignition and spark plugs on the aircraft are usually shielded to eliminate
interference. Another example that will produce EMI between the motor and solenoid
switch is the use shielding, but the amount of miss-out interference noise from the shield
is quite big; thus switching between the motor and the installation of a magnetic filter is
needed to further reduce noise. This may be just a simple filter capacitor and may also con-
sist of capacitors and an HF choke. After the installation of filters, usually there is no need
to take measures to shield between the motor and the solenoid switch.

7.4.1.4  Static Discharger

The purpose of the installation of a static discharger on the aircraft is to release static
charge, which is generated during the flight on the body, reducing interference to radio
equipment. A static discharge device is usually installed in the rear edge of the control sur-
faces, wing tips, and vertical stabilizer surface. It should be noted that a static discharger
will not protect against lightning strikes.

7.4.1.5  Lightning Protection

Lightning protection has been implemented in the shape of the aircraft design. The exter-
nal structure and enclosure of the aircraft should use almost all metal materials; other-
wise, they will not have sufficient thickness to withstand lightning strikes, so the shape
of the outer structure can be said to constitute the plane’s basic lightning protection. The
metal surface is a shield that prevents lightning from damaging the cabin of the aircraft,
and also blocks the electromagnetic energy from entering the aircraft cable. Inside the
aircraft, arresters are installed as some parts are also vulnerable to the effects of lightning,
such as the alternator and antenna tuners of the aircraft. Once an aircraft is struck by light-
ning, it must undergo a comprehensive inspection.

7.4.2  Ground Station Design Protection Measures

7.4.2.1  Airport Requirements
The EME protection zone of a civil airport includes two parts: the regional civilian air-
port civil aviation radio (station) EME protection zone and civilian airfield EME protection
zone.
EME protection regional of civil aviation radio (station) of a civil airport includes the
rectangular range that is occupied by the civil airport runway. Both ends of the runway
centerline and extension cords as a benchmark are extending 500 m, respectively, to the
sides. EME in the civilian airport zone settings prohibits the unauthorized change of its
technical parameters. Civil airport authorities and civil aviation radio station are set using
182 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

units, which should establish civil airport EME protection zone patrol system. When new
facilities or behavior that may affect airport EME (including changes in topography) are
discovered, one should promptly report them to the local civil aviation authorities. In case
of an emergency or special circumstances, one should report directly to the local radio
regulatory agency.

7.4.2.2  NDB Station Requirements

NDB and ADF work together for the determination of the aircraft relative azimuth with
navigation station and guide the aircraft along the intended route of flight, homing,
approach, and landing, which include close NDB, remote NDB in airports, and route
NDB. The close NDB and remote NDB are set in the runway center extended line; the
distance is 900–11,100 m from the landing end of the runway. Route NDB is usually set
below the route turning point and the checkpoint. Thus, the signal coverage area radius
of the close NDB should be not less than 18.5  km, not more than 70  km; signal cover-
age area radius of the remote NDB and route NDB should be within a radius of 150 km.
And the minimum signal strength should be more than 70 μV/m outside latitude 30°
and should be more than 120 μV/m in latitude 30°. It is 9 dB for industrial, scientific,
and medical equipment interference protection and is 15 dB for a variety of other active
interference protection in NDB signal coverage area. The terrain should be flat, open in
the radius range of 100 m from the center of NDB antenna. The minimum separation dis-
tance is allowed between the centers of NDB antenna and all kinds of terrain as shown in
the following table. The communication and power cables near the NDB station should
be buried, with a distance of 150 m from the center of the NDB antenna. There should be
no obstruction outside 50 m in the bottom center of the antenna as a benchmark over a
vertical opening angle of 3° as shown in Figure 7.3.

Terrain name Minimum Separation Distance

Higher than 3 m of trees, buildings, and roads 50
Railways, overhead low-voltage power lines, communication 150
cables, overhead high-voltage transmission lines below 110 kV
Hills, dams 300
Overhead high-voltage transmission lines above 110 kV 500

3° 3°

50 m 50 m

FIGURE 7.3
(See color insert.) No obstruction outside 50 m in the bottom center of the antenna as a benchmark over a verti-
cal opening angle of 3°.
Electromagnetic Interference to Aeronautical Telecommunications 183

7.4.2.3  LOC Station Requirements

The localizer is an integral part of the ILS system and works in conjunction with the air-
craft receiver, which provides the landing alignment guidance information of aircraft.
Localizer antenna array is usually set in the extension line of the runway centerline, and
the distance from the runway end is 180–600 m. The localizer transmitter sets the hori-
zontally polarized sector synthetic field and the signal coverage area: the range of 10° of
the runway centerline is 46.3 km, and the range of about 10°–35° is 31.5 km, as shown in
Figure 7.4, and the minimum signal strength is 40 μV/m in the coverage area.
Because of terrain conditions, the localizer antenna cannot be set in the extended run-
way centerline; the offset settings can be used. Thus, the maximum allowable deviation
deflection angle is 3°, and the distance from the runway centerline must be less than 160 m,
and the bias settings are only limited as ILS of class I, as shown in the bias antenna array
configuration in Figure 7.5.
In the signal coverage area of the localizer, the radio interference protection ratio is 17
dB to FM signal, is 14 dB to industrial, scientific, and medical equipment, and is 20 dB to a
variety of other active interference.
The protected area of the localizer site is denoted by circular and rectangular regions,
and the center of the circle is the localizer antenna; its radius is 75 m. The length of rect-
angle is 300 m to runway end (whichever is greater) from the localizer antenna extending
away along runway centerline; the width is 120 m. If the localizer antenna has radiation
characteristics of single direction, and radiation pattern around the field is better than 26
dB or more, the protected areas are not included in the shaded region in Figure 7.6.
The machine room of the localizer should be set within the range of the antenna array,
which is less than 30° the direction of antenna array. According to the local topography,
road and power conditions can be set on either side of the antenna; the distance between
the antenna center and the room is 60–90 m.

m
5k
31.

35°

35°

46.3 km

FIGURE 7.4
The localizer transmitter makes the horizontally polarized sector synthetic field and the signal coverage area:
the range of 10° of the runway centerline is 46.3 km, and the range of about 10°–35° is 31.5 km.
184 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Max3.0

Max3.0
Runway

350 m+15 m

FIGURE 7.5
Bias antenna array configuration.

15 m

30°
120 m

30°

75 m

FIGURE 7.6
Site-protected area of the localizer.

There should be no obstacle in the protected areas of the localizer, and the beacon power
cables and communication cables should be buried in the protected areas. Vehicles or aircraft
should not be parked in protected areas, and there should be no ground transportation activi-
ties. Even if airfield lighting facilities must be set in the protection zone, the use of nonmetallic
materials should be maximized, and its height and surface area should be as small as possible
to ensure that its impact on radio navigation signals minimized. Thus, there should be no tall
buildings, large metal reflectors, or high-voltage transmission lines in the area that is in the
range of 15 m within the antenna forward range of 20° and within the distance of 3 km.

7.4.2.4  GS Station Requirements

The glide slope is an integral part of the ILS system and works in conjunction with the
aircraft receiver too, which provides the landing vertical guidance information of aircraft.
The glide slope antenna array is usually set within the landing end of the runway off to
the side, which is of minimal impact to flight, and the distance from the runway center
may be 75–200 m, in which 120 m is usually chosen. The distance of glide slope retreat
away from the runway entrance is usually 200–400 m, which can be adjusted in accor-
dance with the following parameters: glide angle, height of baseline data point, terrain
slope, and so on.
Electromagnetic Interference to Aeronautical Telecommunications 185

The glide slope transmitter makes the horizontally polarized sector synthetic field and
the signal coverage area: the antenna forward horizontal range of 8° and the antenna for-
ward vertical range of 0.45θ°−1.75θ° (θ is the slip angle) is 18.5 km, as shown in Figure 7.7,
and the minimum signal strength is 400 μV/m in the coverage area.
In the signal coverage area of the glide slope, the radio interference protection ratio
is 14 dB to industrial, scientific, and medical equipment and is 20 dB for a variety of
other active interference. The protected areas of the glide slope are shown in Figure 7.8.
Region A should not have any obstructions, roads, or crops; the height of weeds should
not exceed 0.3 m. The longitudinal slope of the region should reflect the runway slope,
and the cross slope should be less than 1%, and the area must be flat with no less than
4 cm of height difference. In the region, there should be no parked vehicles, machinery,
or aircraft and should not have ground transportation activities. The power cable and
communication cable should be buried underground through the region. Region B is
600 m front of the glide slope antenna; there should be no railways, highways, dedicated
surround channel of airports, buildings (except localizer room), high-voltage transmis-
sion lines, dams, trees, hills, or other obstacles. The height of the localizer room and the
obstacle outside 600 m should not exceed the limits of the runway end clearance. The
region should be flat ground; uneven terrain height allowable values are related with the
distance to the glide slope antenna and antenna height and other factors, and the formula
is as follows:

GS
Antenna
8°

8°

18.5 km

1.75 θ
GS
Antenna θ
0.45 θ

18.5 km

FIGURE 7.7
The antenna forward horizontal range of 8° and the antenna forward vertical range of 80.45θ°−1.75θ° (θ is the
slip angle) is 18.5 km.
186 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Region C

Region A Region B

Runway

FIGURE 7.8
(See color insert.) The protected areas of the glide slope.

D
Z < 0.0117
N

where
Z are the uneven terrain height allowable values
D is the distance to the glide slope antenna
N are the wavelengths number of sideband antenna height or the wavelength number of
capture the effect middle antenna

Region C should not have rail and road presence (except dedicated surround channel
of airports); there should be no higher buildings, high-voltage transmission lines, dams,
trees, hills, and other obstacles than the airport side clearance restrictions; terrain slope in
the region should not be more than 15%. The glide slope room should be set in the rear or
side rear of the glide slope antenna, and the distance from the glide slope antenna is 2–3 m.

7.4.2.5  MB Station Requirements

A marker beacon (MB) station is part of the ILS; middle and inner marker beacon station,
which is set in the extended runway center line in accordance with the requirements of the
external and the operating frequency, is 75 MHz. MB emitted into the air vertically tapered
synthetic field, and the signal coverage areas are the longitudinal width of 100–200 m in
the height of 20–40 m, the longitudinal width of 200–400 m in the height of 60–80 m, and
the longitudinal width of 400–800 m in the height of 350–600 m.
When MB is influence by the terrain, the radiation pattern will be distorted, caus-
ing deviation mark locations. When MB and NDB are colocated setting, its antenna is
disposed in the extended runway center line, and from NDB antenna 10–30 m. When
ground conditions do not permit the spot, MB antenna can also be mounted directly on
the NDB room roof.
The active interference protection ratio of marker beacon signals is 23 dB within the
coverage area. Outer MB is 6,500–11,100 m from the runway, in which 7,200 m is usually
used, and middle MB is 1050 ± 150 m from the runway, and inner MB is 75–450 m from
the runway. The outer MB and middle MB deviation should not exceed 75 m from the
extended runway centerline, and the internal MB deviation should not exceed 30 m from
the extended runway centerline.
Electromagnetic Interference to Aeronautical Telecommunications 187

R = 30 m

Zone I
Runway 120°
Zone IV Zone II

120°
Zone III

FIGURE 7.9
The protection zone I, II, III, and IV.

Within the protection zones I and III as in Figure 7.9, the benchmark for the ground grid
or MB antenna lowest unit should not be exceeded. Moreover the standard units with-
out directional radio antenna should be outside the room, and the distance from the MB
­station should more than 30 m, with a vertical opening angle of 20° for obstructions. And
within the protection zones II and IV, there should not be exceeded according to network
or MB antenna lowest unit as a benchmark, in addition to standard units without direc-
tional radio antenna outside the room, and within a distance marker beacon station 30 m,
vertical 45° opening angle of the obstacle as in Figure 7.9.

7.4.2.6  DVOR Station Requirements

DVOR operates in the 108–117.975  MHz band and is in cooperation with the aircraft
receiver and provides comprehensive guidance information to guide aircraft along a pre-
determined route (line) flight arrival and departure and approach for aircraft. DVOR has a
reflective grid surrounding the station; the multipath interference and reemitted radiation
to the radio signal can distort the radiation pattern, resulting in channel bend, swing, and
jitter, which in turn affects flight safety.
The DVOR antenna base center is a reference point to the antenna reflector net plane
for plane. There should not be any obstacles beyond the datum height within a radius
of 100 m, and there should be no roads, buildings, dams, hills, trees, and other obstacles
within a radius of 200 m that exceed the height datum. The opening angle relative to the
plane should not exceed 1.5°. Horizontal aperture angle should not exceed 7° within the
radius of 100–200 m vertical. Obstacle vertical opening angle relative to the reference plane
should not exceed 1.5°. Horizontal aperture angle should not exceed 10° within the radius
of 200–300 m. Within a radius of 300 m, it should not exceed the height of the railway, and
the obstacles relative to the vertical plane of the opening angle should not exceed 2.5°. As
a reference point to DVOR antenna base center and as a reference plane to the antenna
reflector net plane, the radius of 200 m height datum should not exceed 35 kV (or more) of
high-voltage transmission lines, and the radius of 500 m of the plane should not exceed the
height of 110 kV (or more) of high-voltage transmission lines. The DVOR signal interference
protection ratio is 17 dB within the coverage area of FM radio, and the industrial, scientific,
and medical equipment interference protection ratio is 14 dB, and other active interference
protection ratio was 20 dB.
188 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

7.4.2.7  DME Station Requirements

The operating band of DME station is 962–1213 MHz, which provides continuous distance
information to aircraft equipment and guides aircraft along the selected route (line) to
flight arrival and departure and approach. Minimum signal field strength within the cov-
erage area of conventional DME is 689 μV/m, and the minimum signal strength outside
13 km from DME station within the coverage area of precision DME is 689 μV/m, and the
lowest signal strength within 13 km from DME station within the coverage area of preci-
sion DME is 3453 μV/m, in the lowest signal field strength at the approach reference point
is 6140 μV/m.
Within the signal coverage area of DME, the protection to a variety of active interference
is 8 dB. If DME is fitted with LOC, GS, or DVOR together, the same EME requirements will
be claimed with the corresponding device. If DME is fitted as a separate desk, the center
point of the antenna will be a reference point, and the center plane of the antenna will be
a reference plane; there should be no obstacle within a radius of 50 m beyond the reference
plane, there should be no obstacles beyond a radius of 50 m vertical plane opening angle
of 3°, which should not have transmission lines of 110 kV and above of high voltage within
500 m.

7.4.2.8  Precision Approach Radar Station Requirements

Precision approach radar transmit horizontal and vertical scanning beams alternately
and receive the reflected echoes of aircraft, mainly as a measuring device for guiding the
aircraft approach and landing. It is usually arranged in flat areas or outside the airport
landing road; the distance from the runway centerline is 120–225 m, retreat away from
the landing point is not less than 915 m, and the angle of the connection point of the
landing runway centerline is formed less than 9°. Operating frequency is 9370  MHz ±
30 MHz.
It should be flat and open around radar stations, and within the coverage area, there
should be no obstacle in less than 500 m from the antenna as a benchmark higher than 0.5°
opening angle.

7.4.3  Study Case

A straight line from the radio test site to an airport is 25 km, and the highest point between
the site and the airport is intermediate 514 m above sea level. The main operating param-
eters of the transmitting device are as follows:

Operating Maximum Transmit

Frequency Power (W) Antenna Form Gain (dBi) Work Type
2–30 MHz 125 Omnidirectional 3 Peer to peer
30–512 MHz 50 Omnidirectional 3 Hopping networking
610–960 MHz 3 Directional 14 Peer to peer
1350–1850 MHz 3W Directional 17 Peer to peer
4.4–5 GHz 20 Directional 21 Peer to peer
5.625–5.825 GHz 2 Directional 21 Peer to peer

These bands are able to work for the frequency range of the device and do not repre-
sent the frequency of tests used to communicate, and which is crossed with air naviga-
tion frequencies in the VHF range. The communication frequency of the radio test site
Electromagnetic Interference to Aeronautical Telecommunications 189

is irrelevant to aeronautical navigation and communication frequencies, so there are no

co-channel interference problems, according to the formula

⎛ 444 × 103 Pt ⋅Gt ⎞ ⎛ 2π ⋅ ht ⋅ hr ⎞

E=⎜ ⎟ × sin ⎜⎝ λ ⋅ d × 103 ⎟⎠
⎝ d ⎠

where
E means the space wave field strength
Pt means the transmitter power
Gt means antenna gain
d means the distance between receive antenna and transmit antenna
ht means the height of the transmit antenna
hr means the height of the receive antenna
λ means the radio wavelengths (λ = c/f)
f means the RF

Protection rate of field strength can be calculated as follows:

Aeronautical Aeronautical
Communications Communication
Lowest Signal Interfering Requirements Actual
Frequency Strength Signal Strength of Protection Protection
Device Type Range (MHz) (dBμv/m) (dBμv/m) Rates (dB) Rates (dB)
LOC 108–111.975 32 2.6 20 29.4
DVOR 108–117.975 39 2.6 20 36.4
GS 328.6–335.4 52 12 20 40
MB 75 63 −0.7 23 63.7

Interference protection distance can be calculated from the formula

⎛ E30 −ES +R ⎞
⎜⎝ 20A ⎟⎠
d = 30 × 10

where
d means the protection distance
E30 means the equipment interference allowed values
ES means the signal strength of protective equipment
R means the protection rate
A means the attenuation coefficient

Interference protection distance can be calculated as follows:

Frequency Device Interference Minimum Distance from Actual Linear

Device Type Range (MHz) Allowed Values dB (μv/m) the Airport Distance
LOC 108–400 40 1687 m (the lowest 25 km
GS frequency)
DVOR
3308 m (the highest
frequency)
190 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Based on the aforementioned analysis and calculation results, the radio test site for field
emission device protection rate to airport navigation equipment is more than the required
value 9 dB, and the protective distance is greater than the required value of 15–20  km.
Thus, the radio test site does meet national standards and industry standards.

References
1. Walen, D. Lightning and HIRF Protection. Washington, DC: Federal Aviation Administration,
1998.
2. National Standardization Management Committee, Electromagnetic environment require-
ments for aeronautical radio navigation stations, GB6364-2013, China Standardization Press,
2013.
3. Walen, D. Aircraft Electromagnetic Compatibility. Washington, DC: Federal Aviation
Administration, 2002.
4. Ely, J. J. Electromagnetic interference to flight navigation and communication systems: New
strategies in the age of wireless. Reston, VA: American Institute of Aeronautics and Astronautics,
2005.
5. NTIS. High-intensity radiated fields (HIRF) risk analysis. Springfield, VA: National Technical
Information Service (NTIS).
8
Analysis and Modeling of the QoS
Mechanism in ATN

Douzhe Li and Zhao Li

CONTENTS
8.1 Introduction......................................................................................................................... 191
8.2 Overview.............................................................................................................................. 192
8.3 QoS in Link Layer (Layer 2).............................................................................................. 193
8.3.1 A Brief Introduction of L-DACS1 Data Link Protocol....................................... 193
8.3.2 QoS and RRM (Radio Resource Management) in L-DACS1............................. 194
8.3.3 Algorithms for Radio Resource Management (RRM)....................................... 196
8.3.3.1 Fair-Share Scheduling............................................................................. 198
8.3.3.2 Randomized User Selection Scheduling.............................................. 199
8.3.3.3 Modified Deficit Round Robin with Fragmentation........................... 199
8.3.3.4 RRM Time Complexity Analysis Methods.......................................... 200
8.4 QoS in Layer 3: DiffServ per Hop Behaviors and Class Definition............................. 200
8.4.1 QoS Support in IPv6............................................................................................... 200
8.4.2 QoS in ATN/IPS Context....................................................................................... 204
8.4.3 ATN/IPS PHBs/CoS............................................................................................... 204
8.4.4 DiffServ Code Point Values................................................................................... 204
8.4.5 Traffic Characterization......................................................................................... 205
8.4.6 QoS Support in Wired to Wireless MIPv6/NEMO............................................ 205
8.5 Basic Knowledge and Mathematical Model of QoS Routing....................................... 207
8.5.1 Model of QoS Routing............................................................................................ 209
8.5.2 Routing Problem Model Description................................................................... 210
8.5.2.1 Unicast Routing Problem........................................................................ 211
8.5.2.2 Multicast Routing Problem..................................................................... 211
8.6 QoS Routing Algorithm for Unicast and Multicast....................................................... 212
8.6.1 Protocols and Algorithms for Unicast Routing.................................................. 212
8.6.2 Protocols and Algorithm for Multicast Routing................................................ 213
8.7 Conclusion........................................................................................................................... 216
References...................................................................................................................................... 216

8.1 Introduction
From the overall point of view, the aeronautical telecommunications network (ATN) is a
mobile cellar wireless communication system that aims at transmitting both real-time and
non-real-time data. Quality of service (QoS) is an important technique that can ensure the

191
192 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

transportation of traffic with special requirements. In the ATN/IPS environment, QoS is a

mandatory requirement that should be deployed since the wireless bandwidth is limited
and is shared by multiusers and many types of services. QoS comprises requirements on
all the aspects of a connection, such as service response time, loss, signal-to-noise ratio,
crosstalk, echo, interrupts, frequency response, loudness levels, and so on.
For a thorough review of QoS and discover its application in ATN, first, in Section 8.2,
we introduce QoS generally and then elaborate on the per hop behaviors (PHB) and class
definition in ATN/IPS. Considering the recently developed L-band data link technology
(i.e., L-DACS1), the QoS support in link layer will be introduced in Section 8.3. Some QoS
protocols that are used in the network layer will be discussed in Section 8.4.
Routing is regarded as one of the key technologies for realizing QoS. By means of math-
ematical theory, QoS routing is often modeled as a nondeterministic polynomial time (NP)
complete problem and often solved by heuristic methods. In Section 8.5, some particular
mathematical model that could be used to achieve the QoS requirements is discussed.
Except well know Unicast problem, considering multicast service such as weather fore-
casting and aeronautical chart updating will be widely used in ATN, we will discuss some
particular multicast routing algorithms in Section 8.6.

8.2 Overview
Network resources (bandwidth, error rate, data transmission speed, etc.) are limited in
any type of communication network, but at the same time, the intrinsic characteristic of
a network is sharing the resource and transport data of multitude of applications, includ-
ing high-quality graphics and delay-sensitive data such as real-time voice. In an ATN/IPS
environment, various types of services (include legacy services that are defined in Doc
97051 and newly defined VOIP service) will simultaneously transmit data by utilizing ATN.
How to utilize the resources effectively and sufficiently is an important problem in
designing and managing a particular network. To solve this problem, we introduce the
definition of QoS. QoS is a set of techniques to analyze and manage network resources
and is the metric of overall network performance, particularly the performance seen by
the users in the network.
Bandwidth-intensive applications stretch network capabilities and resources, but
also complement, add value, and enhance every business process. Networks must pro-
vide secure, predictable, measurable, and sometimes guaranteed services. Achieving
the required QoS by managing the delay, delay variation (jitter), bandwidth, error rates,
throughput, and packet-loss parameters on a network becomes the key point to a success-
ful end-to-end business solution.
As we have seen from Chapter 2, there are four layers in TCP/IP protocol stack infra-
structure. We only consider two layers, one of them is the IP layer, also called Layer 3 (L3);
the other is link layer (we only consider media access control [MAC] sublayer in link layer),
also called Layer 2 (L2). Data transportation between L2 and L3 is dependent on an inter-
face, for example, in L-DACS1 protocol, the interference is named SNDCP (subnetwork
dependent convergence protocol, the network layer adaptation service of the SNDCP shall
provide functions to transfer network protocol data units (N-PDUs) transparently over
L-DACS1 subnetwork). In the ATN environment, L2 and L3 all define QoS requirements
(L2-QoS and L3-QoS), but there are some differences between them. The main difference
Analysis and Modeling of the QoS Mechanism in ATN 193

is, L2-QoS depends on the specific data link protocol and L3-QoS is almost a common
standard and self-contained in IPv6. The detail will be discussed in Section 8.4.
In order to facilitate later explanation, first we will explain the related QoS mechanism
in the link layer (L2-QoS). L-DACS1 is one of the promising candidate data links for future
terrestrial aeronautical communication; understanding the QoS in L-DACS1 can help us to
be more familiar with the L2/L3 joint optimization methods that have appeared in many
papers.

8.3  QoS in Link Layer (Layer 2)

8.3.1 A Brief Introduction of L-DACS1 Data Link Protocol
Action Plan 17 (AP17)2 is an internationally sponsored effort to investigate various possible
candidate technologies which can be used to support future aeronautical communications.
Some of the considered and evaluated technologies operate in the L-band, supporting the
[COCRv2]3 requirements. But none of them can fulfill each requirement perfectly. The
main reason is operational spectrum compatibility, since an existing system in the L-band
will cause interference if there is overlap with another system. However, the analysis of an
existing system and desired features of a new system can help us to develop something
suitable or design a technology to mitigate the interference.
Given these features and the most promising candidates, two technical solutions of
L-band digital aeronautical communications system (LDACS) were developed. Before
the final choice of a suitable data link from these two options, some further evaluation is
required.
One choice, L-DACS1, is a frequency division duplex (FDD) configuration based on
orthogonal frequency division multiplexing (OFDM) technology, it has some advantages,
such as high data rate (up to 2.6Mbit/s) and a more robust multipath channel environ-
ment. Also it has reservation based access control, and can support upper layer advanced
network protocols. L-DACS1 is evolved from broadband aeronautical communications
(B-AMC) and TIA-902 (P34) technology and has a close relationship with them.
Another choice is L-DACS2, a time division duplex (TDD) based system, which is derived
from all-purpose multicarrier aviation communication system (AMACS) technologies.
The modulation of L-DACS2 is binary modulation (e.g, continuous-phase frequency-shift
keying, CPFSK). This type of modulation exists in some commercial systems such as GSM
or universal access transceiver (UAT) system. L-DACS2 can also support QoS require-
ments and upper layer network adaptation. Its bandwidth is narrower than L-DACS1, so
the data rate is lower.
AP17 and SESAR proposed that some activities should be taken before 2020, the dead-
line for a final decision, to further consider the proposed two LDACS options:

• Develop detailed specifications for L-DACS1 and L-DACS2

• Develop and test L-DACS1 and L-DACS2 prototypes
• Assess the overall performance of L-DACS1 and L-DACS2 systems

Specific EUROCONTROL contracts covered the initial activities to develop detailed speci-
fications for the L-DACS1 (Project ID 15.02.04.) and L-DACS2 system.
194 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

The interference of L-DACS is mainly from distance measures equipment (DME),

but also there is other interference such as from Joint Tactical Information Distribution
System (JTIDS) which is based on frequency hopping technology. AP17 suggests that
spectrum compatibility investigations should be conducted in a consistent way, i.e., the
testing environment must be standardized to give a fair conclusion.
Note: Another EUROCONTROL contract has focused on the development of the inter-
ference scenarios to be investigated and the definition of acceptability criteria for each
scenario.

8.3.2  QoS and RRM (Radio Resource Management) in L-DACS1

The L-DACS1 data link protocol has several inner modules (i.e., functional blocks).4,5
These modules are connected by some messages whose transmission action is indicated
by a black double arrow “↔”, as we can see from Figure 8.1. In the actual system, the
message is realized by a PDU (physical data unit), which is a block of bytes (octet) that
each bit has specific meaning. Different modules have different PDU, as described in the
following.

PHY: Physical layer based on OFDM16 that realizes the physical frame construction
and deconstruction, coding, decoding, modulation, and demodulation. The mes-
sage that PHY layer uses is PHY PDU. PHY is out of the scope of this chapter.
MAC: Media access control layer provides an abstraction logical channel to hide the
PHY layer; the message that MAC layer uses is MAC PDU.
DLS: Data link service layer. It may be utilized by the link management entity (LME)
for the conveyance of signaling/management data and the SNDCP for the con-
veyance of SNDCP data PDUs or signaling. The message that DLS layer uses is
DLS PDU.

IP layer (various types of IP packets)

Voice Higher layers

Control SNDCP
(VoIP)

Logical link
DLS VI
control sublayer
LME
Media
MAC
control sublayer

PHY Physical layer

FIGURE 8.1
L-DACS1 protocol stack and its upper IP layer.
Analysis and Modeling of the QoS Mechanism in ATN 195

SNDCP: Network layer protocols are intended to operate over services provided by a
variety of subnetworks and data links. L-DACS1 subnetwork can support several
network layer protocols while providing protocol transparency for the user of the
subnetwork service. The message that SNDCP uses is SN-PDU.
LME: Not having a PDU format definition, LME can be seen as the manager of
L-DACS1. LME controls PHY through interface, and also conducts configuration,
resource management, and mobility management.

We have to keep in mind that all aircraft stations in the same cell are controlled by
a single ground station. Before an aircraft station sends data, it should first request
resources (the request message is included in the multiframe of physical layer) from its
controlled ground station and the ground station will give a time slot in the multiframe
(for the definition of multiframe, please refer to L-DACS1 specification). The resource
allocation is contention free; each aircraft station will have its own static slot in each
transmission.
The Data Link Layer (DLL) provides QoS assurance according to COCRv2 requirements.
Multiplexing of different service classes is possible. Except for the initial aircraft cell-entry,
medium access is deterministic, with predictable performance. Optional support for adap-
tive coding and modulation is provided as well.
The DLS offers its services prioritized, that is, in different classes of service. Service
classes map directly to priorities. The service class of a request shall be used by the ground
station to determine the order and size of resource allocations. Within the data link service
the service class is used to determine the precedence of concurrent service requests. The
classes of service supported by DLS are displayed in Table 8.1.
Service class DLS_CoS_7 shall designate the service class with the highest priority and
DLS_CoS_0 shall designate the service class with the lowest priority. DLS_CoS_7 shall
be reserved for LME signaling. DLS_CoS_6 shall be reserved for the packet mode voice
service (VoIP).
QoS mapping between applications and link technology service classes is one of the
considerations for the scheduling mechanism. On the one side, L-DACS1 supports 8
different service classes, and on the other side, 12 different application level classes of
service are defined.3 The mapping between application class of service (CoS) and L-DACS1
service classes are not studied here, and we only assume two different application CoSs
are mapped to two different L-DACS1 service classes.

TABLE 8.1
DLS CoS Mapping onto ATN/IPS Application Levels
Class of Service Priority Comment
DLS_CoS_7 Highest Reserved for LME Mapping to LME control
DLS_CoS_6 Reserved for packet mode voice signals and VoIP
DLS_CoS_5 Mapping to other
DLS_CoS_4 applications according
DLS_CoS_3 to COCRv2
DLS_CoS_2
DLS_CoS_1
DLS_CoS_0 Lowest
196 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

8.3.3  Algorithms for Radio Resource Management (RRM)

According to L-DACS1 specification, both FL (forward link, means the data transmission
from ground station to aircraft station) and RL (reverse link, means the data transmission
from aircraft station to ground station) resources shall be assigned by the GS’s (ground
station’s) RRM function. Implementation specific details of the resource allocation proce-
dures are not involved in L-DACS1 specification, so several popular RRM algorithms will
be discussed in this section.
For better understanding, the readers should be familiar with the steps of L-DACS1
radio resource request and allocation.
If several aircraft belonged to a ground station, the resource request is periodically sent
by every aircraft station to indicate whether it has data to be transmitted. Ground station
will periodically receive the requests from aircrafts that are under its control, then the
ground station should calculate and allocate available resources (i.e., bandwidth) to each
user (i.e., aircraft). In the RL, L-DACS1 only supports one type of bandwidth request mech-
anism: explicit contention-free bandwidth request and response. An aircraft will have its
privileged time slot for data transmission; if the ground station does not allocate a specific
slot for an aircraft, it cannot transmit data in that period. In this mechanism, each aircraft
explicitly sends its bandwidth request in the dedicated control channels (DCCH) slot in
order to receive resources for the RL transmission to the BS in the upcoming data slot (i.e.,
RL Data).
The IP datagram is first packaged through SNDCP interface to force an SN-PDU. Then,
DLS will separate SN-PDU into several queues according to different CoS. Finally, each
queue will be segmented into several segments to force DLS_PDU. A critical point is the
size of segmentation that is based on the resource allocation that is described as follows.
A resource request is transmitted by a MAC PDU that can be seen in Table 8.2; there are
two types of requests, single resource requests (SRSC_RQST) and multiple resource requests
(MRSC_RQST). Each type of resource request should indicate the CoS type and how
many PHY PDUs of this CoS is requested. The resource request MAC PDU will be con-
verted to a PHY PDU and sent by a logical channel called DCCH; actually, the logical
channel is sited in a physical time slot that is called multiframe (MF). The details of MAC
PDU can be found in Chapter 8 of L-DACS1’s specification.4
After the ground station has received every user’s resource request, it will use some
algorithms to allocate resources. The allocation result will be sent to aircraft by a MAC

TABLE 8.2
Resource Request Format
Field Size (bit) Description
D_TYP 4 Resource request
SC 3 Service class
NRPP 8 Number of RL PHY PDUs
D_TYP 4 Resource request
ENT 3 Number of entries (the number of CoS)
SC1 3 Service class
NRPP1 8 Number of RL PHY PDUs
… … …
SC8 3 Service class
NRPP8 8 Number of RL PHY PDUs
Analysis and Modeling of the QoS Mechanism in ATN 197

TABLE 8.3
Resource Allocation Format
Field Size (Bit) Description
C_TYP 4 RL allocation
SAC 12 Subscriber access code
RPPO 8 RL PHY-PDU offset
NRPP 8 Number of RL PHY PDUs
CMS 3 Coding and modulation scheme
RES 9 Reserved
CRC 4 Cyclic redundancy checksum

PDU, which can be seen in Table 8.3. It indicates the ID of a specific aircraft (subscriber
access code) and the resources (how many PHY PDU can be sent to the aircraft in the next
MF for both RL and FL).
It should be clear that there are two features in L-DACS1 resource allocation:

1. Data queues (aircraft) and scheduler (ground station) are located in separate com-
partments; this is different from other wireless communication data link protocols.
2. Normally in every MF, the DLS entity where the queues are located sends band-
width requests for every user to the LME entity where the scheduler is running.

While a strict priority scheduling mechanism or weighted fair queuing can be applied
among different service classes at the first level according to Table 8.1, another schedul-
ing algorithm is used for each service class at the second level. Three different scheduling
mechanisms are investigated for L-DACS1 by Serkan Ayaz6:

1. Modified deficit round robin with fragmentation (MDRR)

2. Fair-share scheduling (FSS)
3. Randomized user selection scheduling (RANDUS)

Based on the above mentioned schemes, Figure 8.2 shows the main L-DACS1 scheduling
architecture for the RL and FL in L-DACS1. The scheduling algorithm for FL and RL is
running in the LME block and packets are queued in the DLS block.
Strict priority scheduling or weighted fair queuing mechanisms can be used at the first
stage depending on the usage of application types. If currently defined ATN/IPS applica-
tions are only used in the network, then strict priority scheduling is enough since the high
priority message volume is significantly smaller compared to the low priority message
volume. In this case, we do not expect any bandwidth starvation problem for low priority
messages. However, if VoIP with high priority is accepted to be used in the future ATM
then it might cause bandwidth starvation. In this case, usage of a weighted fair queueing
scheduling mechanism is more acceptable since it allocates a certain portion to all service
classes from available bandwidth depending on its priority level.
For this reason, deficit round robin (DRR) is modified in a way that it operates on the
bandwidth request of each user (sent by DLS) and not based on the size of the head-
of-line packet. In the implementation, the quantum size (Qsize) is selected equal to
the packet size. The modified algorithm is able to run for both links (i.e., FL and RL)
separately.
198 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

IP datagrams IP datagrams

L-DACS1 aircraft station L-DACS1 ground station

SNDCP SNDCP

Forward link
DLS Return link
packets (SDUs) packets (SDUS)
LME DLS
Classification Resource Classification
First level
scheduling request from
all CoS for FL
CoS-1 CoS-2 CoS-8 LME CoS-1 CoS-2 CoS-8
… Second level …
scheduling Resource
allocation for
Segmentation FL Segmentation

Return link
frames (PDUs) Forward link
MAC (logical channels) MAC (logical channels) frames (PDUs)
MF MF MF MF
MF MF MF MF

Resource request from all CoS for FL

Resource allocation for FL

CoS: Class of service DLS: Data link services

FL: Forward link LME: Link management entity
MAC: Medium access control PDU: Protocol data unit
SDU: Service data unit SNDCP: Subnetwork dependent convergence protocol

FIGURE 8.2
L-DACS1 link scheduling architecture. (Adapted from Ayaz, S. Mobility and radio resource management in
future aeronautical mobile networks. PhD thesis. Erlangen, Germany: University Erlangen-Nürnberg, 2012)

The author not only describes the algorithms, but also given some performance evalu-
ation. The result is, according to experiments, the tradeoff can be shown between frag-
mented packet transmissions (which is directly proportional to the overhead) and fairness
among different users. On the one side, FSS provided best fairness results while causing
the highest number of fragmented packet transmissions, on the other side RANDUS pro-
vided worst fairness results with the least fragmented packet transmissions. From this
perspective, DRR provided a good balance between these two criteria.
The resource allocation problem can be described as the total bandwidth W that will be
allocated to N users. Ground station knows each user’s situation (how many CoS and how
many PHY PDUs for each CoS). FSS and RANDUS are relatively simple, so we will first
describe them.

8.3.3.1  Fair-Share Scheduling

FSS is designed to allocate resources equally to all users. In our case, when the LME
receives resource requests of all active users, it directly divides the total available band-
width by the number of users and assigns the corresponding portion to each user. There
are two important considerations for the algorithm:

1. In case the resource request of a user is less than the available bandwidth for a
single user then the remaining resources are uniformly distributed among the
other resource requesting users.
Analysis and Modeling of the QoS Mechanism in ATN 199

2. Similarly, in case there are some remaining bytes or PDUs left due to division of
total available resources to the number of resource requesting users, then these
remaining bytes or PDUs are also uniformly distributed among the resource
requesting users.

8.3.3.2  Randomized User Selection Scheduling

The idea of RANDUS algorithm is to assign resources to users randomly considering uni-
form distribution in every MF as shown in Algorithm 6 of Serkan Ayaz’s PhD thesis6. It is
a basic scheduler that does not consider any fairness criteria when distributing resources
among different users. In our implementation, random shuffle function of C++ STL algo-
rithms library is used for randomizing the resource requesting users.

8.3.3.3  Modified Deficit Round Robin with Fragmentation

DRR, also called deficit weighted round robin (DWRR), is a scheduling algorithm for the
network scheduler. DRR is a modified weighted round robin and was proposed by M.
Shreedhar and G. Varghes in 1995. There are two important concepts in DRR we should
know:

• Deficit counter (DC): The number of bytes that a flow is allowed to transmit when
it is its turn
• Quantum size (Qsize): The number of bytes that is added to the deficit counter of
flow in each round (quantum = amount of credit per round)

Applying different quantum to different flows, weighted fairness can be achieved.

Selecting the size of DC and Qsize is dependent on the specific application.
When a packet arrives to the queues, its flow number is checked first. The flow number
can be seen from the Netscape server application programming interface (NSAPI) field in
SN PDU. In case the flow number is in the list (i.e., active list), the packet is just enqueued
corresponding to the specified flow number. In case the flow number is not in the list, it
is inserted in the active list and the packet is enqueued corresponding to the specified
flow number. In our case, when the LME module receives resource requests of all users,
it performs similar operation for the insertion of the user to active list and setting deficit
counter (DC) to zero.
Afterwards, the dequeuing operation of packets takes place while the algorithm is run-
ning according to the original algorithm7 such that it first checks whether the DC is greater
or equal to the size of the head-of-line packet. If this is the case, the algorithm performs the
dequeuing operation, sends the packet, and decreases the DC by packet size. In the other
case, the algorithm just increases the DC by a quantum size (Qsize) and does not send the
packet. Since the algorithm needs to know the size of the head-of-line packet, it requires
certain modification for L-DACS1. The reason is that, in L-DACS1, queues and scheduler
are located in separate compartments.
In L-DACS1, normally in every MF, the DLS entity where the queues are located sends
bandwidth requests for every user to the LME entity where the scheduler is running. For
this reason, we had modified the DRR in a way that it operates on the bandwidth request
of each user (sent by DLS) and not based on the size of the head-of-line packet as shown
in Algorithm 4 of Serkan Ayaz’s PhD thesis6 (i.e., lines 11–51). In our implementation,
200 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

we select the Qsize equal to the packet size. The modified algorithm is able to run for both
links (i.e., FL and RL) separately.
It also needs to be mentioned that the Qsize is selected in terms of bytes on the FL and
in terms of PDUs on the RL according to the bandwidth requests from DLS in BS and DLS
in aircraft, respectively. For real-time services (e.g., VoIP), the quantum size is selected as
103 B on the FL and 8 PDUs (which makes 112 B for the lowest coding modulation scheme)
on the RL. For file transfer on the FL, we select a quantum size of 1091 B. Since the resource
requests of different CoSs are sent separately, it is possible to differentiate which quantum
size is used for which CoS traffic.

8.3.3.4  RRM Time Complexity Analysis Methods

Time complexity analysis for three methods is analyzed in Serkan Ayaz’s PhD thesis.6
It can be found that the time complexity of the scheduling process is correlated with
two components but is independent from the scheduling algorithm. One of the compo-
nents, LME, will receive the resource request messages that are transmitted by DLS. Then
resource schedule algorithms will be used to process different requests. Assume there
are N users in the system that are requesting resources. The algorithm needs to process
each user’s request, therefore the complexity of messages is O(N).6
Another component is the dequeuing operation. After the resources have been
assigned to each user by the algorithm, if the results indicate that M (M ≤ N) users have
been entitled to transmit data, then the data packages corresponding to M users will
be dequeued. Each user’s data packages in DLS should be searched before they can be
dequeued. In a practical system, we can use an algorithm library to realize the operation.
Here, in order to measure the complexity, we take the “find” function of a C++ STL as an
example. The “find” function can be utilized to search data packages for each user, as we
know the complexity of the “find” function is O(log N). If M < N, the total complexity is
O(M) × O(log N), the maximum complexity is O(N log N), if M = N. In FL, resource assign-
ment is by byte, while in RL, resource assignment is through PDU package.
The performance of different scheduling algorithms is based on considering two differ-
ent ATM applications. We restrict our work to scheduling algorithms running at the sec-
ond level and focus on two different service classes separately. In the first class of service,
we consider file transfer of graphical weather information (i.e., WXGRAPH service). In the
second class, we consider VoIP as an example for real-time service. Using the integrated
L-DACS1 model, we thoroughly analyzed the performance of three different scheduling
algorithms for each service class in terms of delay and bandwidth fairness.

8.4  QoS in Layer 3: DiffServ per Hop Behaviors and Class Definition
8.4.1  QoS Support in IPv6
Several features were added to the IPv6 specification in addition to 128-bit addressing as
the IPv6 specification made its way through the IETF committee process. The added fea-
tures include levels of assured service, enhance security, and improved reliability.
QoS is an important term and an emerging feature of modern networks. IPv4 networks
typically give each and every packet a best level of effort service, even if the content of every
packet is not really important or time-sensitive.
Analysis and Modeling of the QoS Mechanism in ATN 201

An IPv4-based system has no way to differentiate between data payloads that are time
sensitive, such as streaming video or audio, and those that are not time-sensitive, such as
status reports and file transfer. Streaming audio and video application are very sensitive
to a delay of a few packets—lips move without sound or picture break up—but IPv4 has
no way to prevent those problems.
If a packet is lost in transit, TCP recognizes the loss and requests a retransmission, but
only after an inevitable delay. The single delayed TCP packet is probably part of a much
larger packet of audio or video data, so the entire big packet is delayed and probably
thrown out because the smallest part did not arrive on time.
IPv6 provides a way for applications to request handling without delay throughout the
wide area network (WAN). The term often used to describe this is low latency. Streaming
audio and video requires low latency through high priority. To prevent a breakdown in the
scheme, various applications can share a connection via priority level.

• Level 0—No specify priority

• Level 1—Background traffic (news)
• Level 2—Unattended data transfer (email)
• Level 3—Reserved
• Level 4—Attended bulk transfer (FTP)
• Level 5—Reserved
• Level 6—Interactive traffic (Telnet, Windowing)
• Level 7—Control traffic (routing, network management)

Packet fragmentation is a major source of packet delay, or high latency, under IPv4. Each
device attached to a network has a payload data limit set inside the Ethernet packet. If a
program is generating streaming data, such as video or audio, the data stream will be split
up into a string of packets, each carrying the maximum payload.
With different devices, these payload sizes are set differently and it is possible that
between the originating source and the destination, some transmission path, particularly
an asynchronous transfer mode (ATM) link, will have a smaller payload size.
The ATM equipment will chop the already fragmented stream of data into even smaller
pieces. An ATM switch could divide the data carried in one single Ethernet packet into 20
ATM cells. Somewhere in the dividing and rebuilding of all the data, it is likely that a cell
or a packet will be dropped or delayed.
IPv6 uses a more sophisticated approach to handle data from programs requesting pri-
ority handling. The originating device will query the destination in order to determine
the maximum size of the payload that can be handled across the complete route. Then it
adjusts its own parameters and will not load the originating packets with more data than
the network can handle.
This approach reduces fragmentation and latency but can also result in inefficient utili-
zation. The tradeoff is that with shorter payloads, it will achieve a higher bandwidth with
prompt arrival.
QoS functionality will have to be included on every networked device in order to be
implemented. Unavailability of functionality on certain devices will cause it to fallback to
a standard handling with just an additional layer to pass through.
Currently IPv6 provides support for QoS marking via a field in the IPv6 header.
Similar to the type of service (ToS) field in the IPv4 header, the traffic class field (8 bits) is
202 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 8.4
IPv6 Packet Header
Version = 6 Traffic Class Flow Label
Payload length Next header Hop limit
Source address
Destination address

available for use by originating nodes and/or forwarding routers to identify and distin-
guish between different classes or priorities of IPv6 packets (Table 8.4).
The traffic class field may be used to set specific precedence of differentiated services
code point (DSCP) values. These values are used in the exact same way as in IPv4.
The key advantage with the flow label is that the transit routers do not have to open the
inner packet to identity the flow, which aids with the identification of the flow when using
encryption and other scenarios. In the ATN environment, IPS nodes shall set the flow label
field of the IPv6 header to zero, as it is not used in the ATN/IPS.
IPv6 also has a 20-bit field known as the flow label field (RFC 3697). The flow label enables
per-flow processing for differentiation at the IP layer. It can be used for special sender
requests and is set by the source node. The flow label must not be modified by an interme-
diate node.
L3-QoS uses ToS (type of service) and deals with IP address like

1. Classification/prioritization of packets in forwarding path based on DSCP IP

header field
2. Policy and allocation of priorities along the transmission path

Each router supporting DSCP needs to be configured accordingly (priorities).

The routers along a path (source to destination) do not store the state about the flow (e.g.,
number of packets already transmitted). Instead the routers apply a policy on each packet
individually (Figure 8.3).

• Classifier: Classifies packet into an internal class. For intermediate routers, the
classifier may be missing (packet already classified by AS ingress router)

Meter Packet queues

Shaper
IP packet
IP packet Classifier Marker dropper
(scheduler)

FIGURE 8.3
General logical architecture of a router.
Analysis and Modeling of the QoS Mechanism in ATN 203

• Marker: Measures the temporal properties of the packet stream selected by the
classifier and instructs the marker and shaper/scheduler to treat a specific packet
accordingly (e.g., drop a packet that is out of profile, i.e., the stream already used
100% of the available bandwidth). Marks packets according to the class defined by
the classifier.
• Shaper/scheduler: Extracts packets from queues according to a local policy and
sends the packet.

ToS field was too inflexible and redesigned to a single field named DSCP.
DSCP contains a number that indicates the PHB to be applied to the IP packet
(Figure 8.4).
Packets are classified (and DSCP field marked) at the ingress into a domain (e.g., autono-
mous system or administrator domain in ATN/IPS).
Intermediate routers in domain B prioritize packets according to the DSCP field in IP
header.
Domain B egress router shapes and schedules packets. The process can be seen in
Figure 8.5.

6 bits 2 bits

DSCP CU

DSCP—DiffServ code point

CU—currently unused

FIGURE 8.4
New definition for IPv6 QoS label in RFC2474.

Administration DiffServ Administration DiffServ Administration DiffServ

domain A domain B domain C

2 4 1 3 4 3 2 1

Service classes

1 Premium

2 VoIP
1 1 CoS1
3 Weather forecast
Ingress node: 2 2 CoS2 Egress node:
Marking (classification) 4 Normal message
Queueing and
at ingress point 3 CoS3 shaping
4
CoS4

FIGURE 8.5
Illustration of DiffServ domain effect.
204 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

8.4.2  QoS in ATN/IPS Context

ATN/IPS communication service providers are likely to make use of the same IPS infra-
structure for ATN and other non-ATN defined applications; for example, Air Traffic Service
Message Handling System (ATSMHS) and surveillance data. Sharing of resources can be
at different levels, ATN/IPS applications can use the same type of CoS as non-ATN appli-
cations over the same IP-routed infrastructure. Alternatively, ATN/IPS communication
service providers may only wish to share the same physical infrastructure and operate a
VPN per service; in this case a separate CoS model can be applied to each virtual private
network (VPN) service, one being the ATN/IPS. Fundamentally, ATN/IPS communica-
tion service providers have flexibility in how they enable CoS for the ATN/IPS over their
infrastructure.
For CoS definitions, it is essential that ATN/IPS traffic is sufficiently qualified in order to
properly mark ingress traffic. As the IP packet enters the network core, PHBs are enforced,
depending on the packet marking. ATN/IPS communication service providers will need
to handle unmarked or premarked ingress traffic and be prepared to mark or remark the
traffic before it is routed over their infrastructure. The internal techniques, mechanisms,
and policies to enforce the PHB within the communications service provider networks are
considered out of the scope of the ATN/IPS.

8.4.3  ATN/IPS PHBs/CoS

The ATN/IPS is to support legacy ATN applications over the IPS. Currently, this intended
support covers CM(DLIC), FIS(ATIS), CPDLC, ADS-C, and ATSMHS. Indeed, DIR is only
specified for ATN/OSI and it is foreseen that AIDC will be implemented through regional
solutions. As each ATN application is mapped to a given CoS, the dynamic support of dif-
ferent priorities per user message category is not considered.
Table 8.5 provides an example of an administrative domain that supports several appli-
cations and CoSs labelled very high, high, normal, and best effort.
In order to mark ingress traffic, the ATN/IPS provider has several means to identify the
traffic: destination transport port number, IP source address, and IP destination address.
ATN/IPS is not recommended to use the DSCP/ToS value set by the application or prior
communication service provider, because the DSCP/ToS field may be incorrectly config-
ured or unknown, and will not be the optimum approach.

8.4.4  DiffServ Code Point Values

PHB is indicated by encoding a 6-bit value—called the DSCP—into the 8-bit differentiated
services (DS) field of the IP packet header. The DSCP value of the field is treated as a table
index to select a particular packet-handling mechanism. This mapping must be configu-
rable and administrative domains may choose different values when mapping code points
to PHBs. However, it is widely accepted that DSCP value 101110 refers to EF (expedited
forwarding).
Table 8.6 provides an example of mapping DSCP values to ATN/IPS PHBs where a num-
ber of applications share the same IP network infrastructure. In this table, air–ground
applications have been assigned with the special class selector code points as specified
in Document 9880 for the ATN IP SNDCF, but within the ATN/IPS it would be better to
make use of AF PHBs to avoid any interaction with legacy applications that make use of
IP precedence.
Analysis and Modeling of the QoS Mechanism in ATN 205

TABLE 8.5
ATN/IPS Priority Mapping to Classes
Priority/Application Mapping Traffic Identification (Ingress)
Class Drop ATN TCP/UDP
(CoS Type) Precedence ATN Priority Application Port IP Address
Very high (EF) Voice (VoIP) RTP numbers —
16384-32767
High (AF) 1 0 — — —
1 — — —
2 — — —
3 ADS-C TCP 5913 The source or destination
UDP 5913 address will be part of a
CPDLC TCP 5911 reserved address space
UDP 5911 assigned to mobile service
providers
Normal (AF) 1 4 AIDC TCP 8500a
FIS(ATIS) TCP 5912 The source or destination
UDP 5912 address will be part of a
reserved address space
assigned to mobile service
providers
2 5 METAR — —
3 6 CM(DLIC) TCP 5910 The source or destination
UDP 5910 address will be part of a
reserved address space
assigned to mobile service
providers
ATSMHS TCP 102
7
Best effort 8–14 — —
(default)
Source: ICAO, Manual for the ATN Using IPS Standards and Protocols, Doc 9896, Draft Version 21, 2013.
a This is applicable when OLDI/FMTP is used as a means to enable AIDC services.

8.4.5  Traffic Characterization

Traffic characterization is a means to express the type of traffic patterns, integrity, and
delay requirements. It provides further information to the communication service provider
in order to fully meet the user requirements within a specific network operation. Typically,
traffic characterization information is part of the contracted service level agreement in
which further parameters are defined such as service delivery points, service resilience,
required bursting in excess of committed bandwidth, service metric points, mean time to
restore (MTTR), and port speeds.
Table 8.7 provides an example of traffic characterization for ground–ground services,
which is derived from the Pan-European Network Services (PENS) specifications.

8.4.6  QoS Support in Wired to Wireless MIPv6/NEMO

The above sections only consider the QoS problem in L2 and L3 separately; it is obvi-
ous that ATN/IPS is a complex and integrated mobile system and handover between two
206 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

TABLE 8.6
Example of DSCP to PHB Mapping
DSCP Value PHB Application
000000 CS0 Best effort
001000 CS1
001010 AF11 AIDC
001100 AF12
001110 AF13
010000 CS2 CM
010010 AF21 ATSMHS
010100 AF22
010110 AF23
011000 CS3 FIS
011010 AF31 Voice recording
011100 AF32
011110 AF33
100000 CS4 CPDLC, ADS-C
100010 AF41 Voice signaling
100100 AF42
100110 AF43
101000 CS5
101110 EF Voice
110000 NC1/CS6
111000 NC2/CS7

TABLE 8.7
Example of Traffic Characterization
ATN Average Expressed Typical Bandwidth Network
Application Message Length Integrity Jitter (Point to Point) (kbps) Delay (1-way)
Voice (VoIP 70 bytes — <15 ms 12 <100 ms
using G.729)
OLDI/FMTP 150 bytes 1 user corrupt N/A 10 <1 s
(regional AIDC) message in 2000
ATSMHS 3 kb 10−6 (in terms of N/A 20 <5 s
1000 bytes
message blocks)

ground stations is a common problem. How to keep the QoS (especially required by time-
sensitive applications) during handover is attractive and challenging. In Chapter 3, we
have introduced MIPv6 and four types of extensions such as PMIPv6, HMIPv6, FMIPv6,
and NEMO. The purpose of these extensions is reducing the signaling time to expedite
the data rate.
QoS performance in mobile environment depends on two aspects: The first is how the
node mobility affects end-to-end QoS guarantees and the second is how to apply the exist-
ing QoS technologies in wired networks to wireless networks, that is, how to append mobil-
ity support to these solutions and how these solutions suit the wireless link characteristics.
Analysis and Modeling of the QoS Mechanism in ATN 207

As mentioned before, QoS is closely related to L2 and L3, so most of the study is to con-
sider a hierarchical method (or cross-layer method, their meanings are almost the same)
to solve this problem. We will provide some research achievement in this section to find a
reasonable approach.
Lee et  al. proposed a cross-layer hierarchical network mobility framework called
Hi-NEMO for all-IP networks.8 The advantage of Hi-NEMO is that it does not have trian-
gular routing between an MNN and the CN. The design is resilient to error-prone trans-
mission, and protocol supports fast QoS provisioning in the NEMO service domain. A
QoS-incorporated handover (QoS-handover) is suitable to fulfill QoS requirements of real-
time multimedia applications on a high-velocity vehicle and not just the reestablishment
of network connectivity.
Kan et al.9 proposed a new two-plane two-tier QoS architecture based on the advantages
of IntServ and DiffServ. This chapter describes how the architecture guarantees an end-to-
end QoS. Finally, the existing mobile IPv6 signaling such as binding update (BU), binding
request (BR), and binding acknowledge (BA) is extended for QoS negotiation and advance
resource reservation.
Ayaz gave the method to speed up the handover performance in an MIPv6 environment.6
Kong et al. have already proven by experiment that PMIPv6 is a promising candidate
solution for realizing the next-generation all-IP mobile networks. As mentioned in Section
2.5.3, PMIPv6 is using LMA and MAG to construct an enhanced function to extend MIPv6.
Liebsch et al.10 proposed a new mobility option, the QoS option, for PMIPv6.
Although some of the studies do not focus on L-DACS1 and ATN/IPS environment,
the concepts can be applied to ATN/IPS environment since ATN/IPS is also an all-IP
network.

8.5  Basic Knowledge and Mathematical Model of QoS Routing

In RFC2386, QoS-based routing is defined as “a routing mechanism under which paths for
flows are determined based on some knowledge of resource availability in the network as
well as the QoS requirement of the flows.” In short, it is a dynamic routing scheme with
QoS consideration.
Before we introduce the QoS routing problem, two concepts that are correlated with
ATN/IPS routing should be first described. For better understanding, an illustration is
also shown in Figure 8.5

1.
Administrative domain: From the view point of administrators, the ATN/IPS inter-
network is composed of several interconnected administrative domains. An
administrative domain can be an individual state, a group of States (e.g., an ICAO
Region), an air communications service provider (ACSP), an air navigation service
provider (ANSP), or any other organizational entity that manages ATN/IPS net-
work resources and services (Figure 8.6).
2.
Autonomous systems (AS): From a routing perspective, interdomain routing pro-
tocols are used to exchange routing information between ASs, where an AS is
a connected group of one or more IP address prefixes. The routing information
exchanged includes IP address prefixes of differing lengths. For example, an IP
208 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Flying through two ASs

BGP4

Autonomous system #1 Edge router Autonomous system #2

FIGURE 8.6
(See color insert.) ATN/IPS routing infrastructure.

address prefix exchanged between ICAO regions may have a shorter length than
an IP address prefix exchanged between individual states within a particular
region.

The ATN/IPS standard only specified the interdomain routing protocol (i.e., border gate-
way protocol (BGP-4) as specified in RFC 4271). The routing protocol within an AS is
a local matter determined by the managing organization. Interdomain routing proto-
cols can be used to exchange routing information among ASs. A single administrative
domain may be responsible for the management of several ASs. BGP is a type of best-
effort routing protocol; it will use the shortest path to the destination (shortest path does
not necessarily mean the path with shortest physical distance. It may also mean the
path with the least cost or fewest hop counts, for instance). In other words, they usu-
ally use single objective optimization algorithms that consider only one metric (band-
width, hop count, and cost). Thus, all traffic is routed to the shortest path. Even if some
alternate paths exist, they are not used as long as they are not the shortest ones. One
drawback of this scheme is that it may lead to the congestion of some links, while some
other links are not fully used.
ATN/IPS is expected to support a wide range of communication-intensive, real-time
multimedia applications or non-real-time applications such as file transfer. The require-
ment for timely delivery of digitized audio–visual information raises new challenges. In
order to ensure QoS performance, one of the key issues is solving QoS routing problem.
It finds the best network routes with sufficient resources for the requested QoS parameters.
The goal of routing solutions is twofold:

• Satisfying the QoS requirements for every admitted connection

• Achieving the global efficiency in resource utilization
Analysis and Modeling of the QoS Mechanism in ATN 209

There are two types of information transportation in ATN/IPS,2 unicast and multicast. Unicast
is a point-to-point transfer that sending messages to a single network destination identified
by a unique address and multicast is group communication where information is addressed
to a group of destination computers simultaneously. In the ATN, the need of sending the
same information to multiple receivers is one of the main requirements of surveillance data
distribution. Using a mathematical model can extract the essence and give a better solution.

8.5.1  Model of QoS Routing

A network can be modeled as a graph that is used to describe and calculate routing prob-
lems. A graph can be written as G(V, E); nodes (V) of the graph represent switches, routers,
and hosts. Edges (E) represent communication links. Since the bidirectional and asymmet-
ric characteristic of most communication links, every link is represented by two directed
edges in the opposite directions. The purpose of solving QoS routing can be described as

1. For unicast, finding a network path that meets the requirement of a connection
between two end users
2. For multicast, finding a multicast tree rooted at a sender and the tree covers all
receivers with every internal path from the sender to a receiver satisfying the
requirement

Routing consists of two basic tasks. First, how to measure and collect network state infor-
mation; second, how to compute routes based on the information collected. The perfor-
mance of any routing algorithm directly depends on how well the first task is solved.
Figure 8.7 gives a brief visualization of the graph model in QoS routing; m and D is the
information source and destination and a–d are nodes included in the network. The link
state consists of residual bandwidth, delay, and cost. Each node should have a state. The
node state can be either measured independently or combined into the state of the adjacent
links. For the latter case, the residual bandwidth is minimally the link bandwidth and the
CPU bandwidth; the delay of a link consists of the link propagation delay and the queue-
ing delay at the node; the cost of a link is determined by the total resource consumption at
the link and the node, such as charges of service providers.

Link state = (bandwidth, delay, cost)

(2, 3, 1)
a b
m

(1, 2, 1) (2, 2, 1)
W
ire
les

(0.1, 2, 3)
s

(2, 2, 5)
Ground G D
Air traffic
station
surveillance center
(2, 2, 1)
(4, 0.5, 6)

c d
(2, 1, 3)

FIGURE 8.7
Abstract network state.
210 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

According to Figure 8.7, the unicast routing problem is defined as given a source node m,
a destination node D, a set of QoS constraints C, and possibly an optimization goal, find
the best feasible path from m to D, which satisfies C. The multicast routing problem is
defined as given a source node m, a set of destination nodes T, a set of constraints C, and
possibly an optimization goal, find the best feasible tree covering S and all nodes in T,
which satisfies C. The two classes of routing problems are closely related. The multicast
routing can be viewed as a generalization of the unicast routing in many cases. These two
problem classes can be further partitioned into subclasses.
Mainly, there are three types of QoS-based routing algorithms, which are called hop-by-
hop routing (also called distributed routing), source routing (also called path addressing)
and hierarchical routing. “They are classified according to the way how the state informa-
tion is maintained and how the search of feasible paths is carried out.”11

• For hop-by-hop routing, the routing table list in each router only contains the
address of the next device towards the destination. Therefore a router only for-
wards a data package according to the next-hop router. This method is the fun-
damental characteristic of the IP internet layer, and is utilized by the well known
Routing Information Protocol (RIP) and most current Internet routing protocols.
• For source routing, the sender of package has partial or complete knowledge about
the network, i.e., the global network state information can be obtained, so the
source router can locally specify the path that the package will be used to reach
the destination router. The advantage of this method is it is easy for troubleshoot-
ing and enhancing the trace-route. The disadvantage is the source does not have
the capability to avoid congestion.
• Hierarchical routing is correspondence to flat routing (e.g. source routing) and is
based on hierarchical addressing. That means the routers are arranged in a hierar-
chical manner, i.e., the actual routers are placed in the bottom level, and they can
be grouped into several next layer nodes. Continuing this process recursively, we
can build a large network. For example, most TCP/IP routing is a two-level hierar-
chical routing, i.e., each IP address consists of two parts, the network portion and
the host portion. The gateway divides a package according to the network portion.
An advantage of hierarchical routing is that it reduces the routing table size from
n2 to log n, but the overhead is increased.

8.5.2  Routing Problem Model Description

ATN/IPS is a wireless network that needs QoS guarantees. However, the dynamic nature
of ATN makes it more difficult to provide QoS, because it is hard to keep routing state
information up to date. One unique problem of wireless network is the mobility of the
nodes, which could lead to the breaking of existing paths and the adding of new links.
Handoff problem happens when an aircraft in the network moves from one ground station
cell to another cell.
One reasonable way to handle this problem is to divide the network links into two types:
stationary links and transient links. Stationary links are those between the stationary
nodes or slowly moving nodes, which are likely to exist for a long time. While transient
links are those between nodes moving very fast. When an existing path is broken, the
traffic flow is rerouted to another feasible path. During the period after the old path is bro-
ken and before the new path is set up, best-effort routing is used to route the traffic flow.
Analysis and Modeling of the QoS Mechanism in ATN 211

For this reason, wireless network normally can only provide soft QoS, which means the
required QoS is not guaranteed for some transient time periods, when the routing path is
broken or the network is partitioned due to the moving of network nodes.

8.5.2.1  Unicast Routing Problem

As we can see from Figure 8.7, if a mobile node sends packages from m to D, Each edge
(e.g., a to b in Figure 8.7, or more general, u to v), is represented by the link between two
vertices e = (u, v) and has associated q weights corresponding to QoS metrics such that
Wi(u, v) ≥ 0 and i = 1, 2,…, q. The constraint for each QoS metric is Li.13 The multiconstrained
path (MCP) problem is to find a path P from a source m to a destination D such that all the
QoS constraints are met, as depicted in the following equation:

wi (P) ≤ Li , i = 1, 2,…,q

The paths that satisfy these constraints are called feasible paths. The solution of the MCP
problem requires a path computation algorithm that finds paths that satisfy all the con-
straints as expressed in the equation. Since the optimal solution of this type of problem for
multiple additive and independent metrics is NP-complete, usually heuristics or approxi-
mation algorithms are used.

8.5.2.2  Multicast Routing Problem

As we know, the mathematical model of a network is a directed graph G = (V, E) with n
nodes (|V| = n) and l links (|E| = l),14 where V is a set of nodes and E is a set of links, respec-
tively. Each link e = (i, j) ∈ E is associated with two parameters, namely, a link cost c(e) ≥ 0
and a link delay d(e) ≥ 0. The link cost, c(e), can be associated with the utilization of the link.
A higher utilization is represented by a higher link cost. The link delay, d(e), is the sum of
the perceived queuing delay, transmission delay, and propagation delay. We define a path
as a sequence of links such that (u, i), (i, j),…, (k, v), belongs to E.
Let P(u, v) = {(u, i), (i, j), …, (k, v)} denote the path from node u to node v. If all nodes u,
i, j,…, k, v are distinct, then we say that it is a simple directed path. We define the length
of the path P(u, v), denoted by n(P(u, v)), as the number of links in P(u, v). A given source
node s ∈ V and destination node d ∈ v, (2s → d, ∞) is the set of all possible paths from s to d.
(2s → d, ∞) = {Pk(s, d)|{all possible paths from s to d; s, d ∈ V, ∀ k ∈ Λ}, where Λ is an index set.
The path cost of Pk is given by φc (Pk ) = ∑ c(e), and similarly, the path delay of Pk is given
e∈Pk

by φD (Pk ) = ∑ e∈Pk
d(e),∀Pk ∈(2 s→d
,∞),(2 s→d
, Δ) and is the set of paths from s to d for which
the end-to-end delay is bounded by Δ. Therefore, (2s → d, Δ) ⊆ (2s → d, ∞).
For the multicast communications, messages need to be delivered to all receivers in the
set M⊆V\{s}, which is called the multicast group, where |M| = m. The path traversed by
messages from the source s to a multicast receiver, mi, is given by P(s, mi). Thus, the multi-
cast tree can be defined as T(s, M) = ∪mi ∈M P(s,mi ) and the messages are sent from s to M
through T(s, M). The tree cost of tree T(s, M) is φc (T(s, M)) = ∑
c(e) and the tree delay is
( )
c∈T
φD (T(s, M)) = max φD ( P(s,mi ))|∀mi ∈M .
In order to measure the complexity of different QoS routing algorithms, often called a
routing decision problem is NP-complete when it is both in NP and NP-hard. The set of
212 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

NP-complete problems is often denoted by NP-C or NPC. NP-complete problems are often
addressed using heuristic methods and approximation algorithms.

8.6  QoS Routing Algorithm for Unicast and Multicast

In the previous section we gave the mathematical model of the QoS routing problem that
had been carefully reviewed by some researchers.12 In this section, first some classical
algorithms will be discussed, and then we discuss some research on algorithms that was
mainly published from 2010 to 2014.

8.6.1  Protocols and Algorithms for Unicast Routing

The unicast information transferred by ATN/IPS can be categorized into two classes,
that is, ground–ground application and air–ground application. Only surveillance data
(i.e., automatic dependent surveillance, ADS) is multicast; all other applications can be
treated as unicast. There are three kinds of routing protocols available to route unicast
packets.

1.
Distance vector routing protocol: Distance vector is a simple routing protocol that
bases its routing decision on the number of hops between source and destination.
A route with the least number of hops is considered the best route. Every router
advertises its best routes to other routers. Ultimately, all routers build up their
network topology based on the advertisements of their peer routers. Example: RIP,
currently is RIPv2, described in RFC2453.
2.
Link state routing protocol: A slightly complicated protocol than distance vector. It
takes into account the states of links of all the routers in a network. This technique
helps routes build a common graph of the entire network. All routers then calcu-
late their best path for routing purposes. Examples: open shortest path first (OSPF)
and intermediate system to intermediate system (IS–IS).
3.
Path vector routing protocol: It will maintain the path information that gets updated
dynamically. Updates that have looped through the network and returned to the
same node are easily detected and discarded. It is different from the distance vector
routing protocol and link state routing protocol. Each entry in the routing table con-
tains the destination network, the next router, and the path to reach the destina-
tion. Example: BGPv4 that was selected in ATN/IPS as the interdomain protocol.

It should be noted that the above classification method of QoS routing is not a conflict with
the classification method in Section 8.6.1 (source routing/hop-by-hop routing/hierarchical
routing). The difference between these two classification methods are the different per-
spectives of QoS routing. Actually, the research interest of unicast QoS routing is mainly
concentrated on link state routing protocol that considers more state information, including
bandwidth, delay, jitter, and so on.
The most commonly used and classical shortest path algorithms are Dijkstra’s algo-
rithm, Bellman–Ford algorithm, and Floyd–Warshall algorithm; many extended algo-
rithms are based on them. The details of these algorithms are carefully discussed
Analysis and Modeling of the QoS Mechanism in ATN 213

in Chapter 2 of Medhi’s book.14 Chen et al.11 and Curado et al.12 have given a thorough
review of unicast algorithms.

8.6.2  Protocols and Algorithm for Multicast Routing

Some multicast routing protocols that were used to exchange multicast routing informa-
tion had already been used in a practical system. These protocols are based on the network
model and some related algorithms. In this section, some existing multicast protocols will
be presented and also the algorithms will be discussed. As the ATN/IPS standard did not
specify the intradomain protocol, it is necessary to understand some of these protocols to
design and analyze an ATN/IPS network.
We introduce three commonly used multicast routing protocols, namely, DVMRP (dis-
tance vector multicast routing protocol), MOSPF (multicast extensions to OSPF) protocol,
and PIM-DM (protocol independent multicast-dense mode) protocol. They are very effi-
cient in environments where multicast group members are densely distributed over the
network.

1.
DVMRP: Derived from the RIP protocol, DVMRP is is the oldest routing protocol
that supports multicast. The router generates a routing table with the multicast
group. The routing table has knowledge of corresponding distances (i.e., num-
ber of devices/routers between the router and the destination). When a multicast
packet arrives at a router, the router will forward it according to the interfaces
specified in the routing table, i.e., delivering tree.
Delivery trees are constructed according to the information on the previous hop
back to the source. First, the router will forward the received packets via each
interface except the one at which the packet arrived at. Second, if a router receives
the packet transmitted by the source router in the first step, but does not want it
to be part of this multicast group, it sends a “prune message” back to the source.
These two steps are called reverse flooding technique.
2.
MOSPF: Defined in RFC 1584 as built on top of OSPF (Version 2, RFC 1583), MOSPF
extended the OSPF protocol. The routers that support MOSPF will maintain a table
of the network topology. This table is constructed via the unicast OSPF link-state
routing protocol. Both MOSFP and DVMRP do not support tunnels.
First, the source router transmits an OSPF link state information advertisement,
then based on the information, the source router can create a distribution tree to
identify each multicast source and group, and another tree for active sources send-
ing to the group. The source and all destination group members should reside in
the same OSPF area, or the whole autonomous system is a single OSPF area. Note
here the route advertisement is realized by Internet Group Management Protocol
(IGMP).
3.
PIM-DM: Developed by the interdomain multicast routing (IDMR) working group
of the IETF. PIM-DM requires the existence of a unicast routing protocol to build
a routing table. It can be treated as the follower of DVMRP protocol. The multi-
cast date from the source router will “flood” to all parts of the network. The last-
hop router will “prune” back to the source if there is no destination group. Dense
mode is more efficient when the communication bandwidth is adequate and the
group members are densely located among regions.
214 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Some algorithms from simple to complex are discussed as follows:

1. The simplest method is the flooding algorithm, which has been already used in
protocols such as OSPF, is the simplest technique for delivering the multicast data-
gram to the routers of an internetwork. When receiving a multicast packet, the
router will first check whether it has seen this particular packet before or it is the
first time that packet has reached this router. If it is the first time, the router will
forward the packet on all interfaces, except the one from which the packet has been
received. Otherwise, the router will simply discard the packet. This way we make
sure that all routers in the internetwork will receive at least one copy of the packet.
This type of setup creates unnecessary traffic and wastes precious network
resources. Each host machine has to process the packet as it arrives on the inter-
face card, thus, wasting CPU cycles. If the volume of the multicast stream is high
enough, it could potentially cause other relevant and control traffic to be dropped.
2. Spanning tree: A better algorithm than flooding is the spanning tree algorithm. This
algorithm that has been already used by IEEE-802 MAC bridges is powerful and
easy to implement. In this algorithm, a subset of internetwork links is selected to
define a tree structure (loop-less graph, it contains no more cycles) such that there
is only one active path between any two routers. Since this tree spans to all nodes
in the internetwork it is called spanning tree. Whenever a router receives a multi-
cast packet, it forwards the packet on all the links that belong to the spanning tree
except the one on which the packet has arrived, guaranteeing that the multicast
packet reaches all the routers in the internetwork. Obviously, the only informa-
tion a router needs to keep is a Boolean variable per network interface, indicating
whether the link belongs to the spanning tree or not.
3. The reverse path broadcasting (RPB) algorithm: This is a modification of the span-
ning tree algorithm. Instead of building a network-range spanning tree, an implicit
spanning tree is constructed for each source. Based on this algorithm, whenever a
router receives a multicast packet on link “L” and from source “S”, the router will
check and see if the link “L” belongs to the shortest path toward S. If this is the case,
the packet is forwarded on all links except L. Otherwise, the packet is discarded.
Three multicast trees from two sources of the test network are shown in Figure 8.8.
This algorithm is efficient and easy to implement. Furthermore, since the packets
are forwarded through the shortest path from the source to the destination nodes,
it is very fast. The RPB algorithm does not need any mechanism to stop the for-
warding process. The routers do not need to know about the entire spanning tree
and since the packets are delivered through different spanning trees (and not a
unique spanning tree), traffic is distributed over multiple tress and network is
better utilized. Nevertheless, the RPB algorithm suffer from a major deficiency: it
does not take into account the information about multicast group membership for
constructing the distribution trees.
There are other two algorithms that belong to RPB family, TRPB (truncated reverse
path broadcasting) and RPM (reverse path multicasting).

In recent published papers, many intelligent algorithms were proposed such as genetic,
particle swarm, and simulated annealing algorithm.17,18 We only cite some papers that pub-
lished in some relatively important journals.
Analysis and Modeling of the QoS Mechanism in ATN 215

1
A B
2
3 4 C

6 5
D E

RPB tree from source A

1
A B
2
3 4 C

6 5
D E

RPB tree from source C

FIGURE 8.8
Reverse path broadcasting algorithm.

Leela20 introduced a genetic algorithm (GA)–based multicast algorithm that satisfies

the multiple constraints of multimedia applications. A heuristic called multiconstraint
QoS unicast routing using genetic algorithm (MURUGA), which incorporates multiple
constraints required by multimedia applications to find the feasible path satisfying the
constraint requirement has been proposed, designed, and simulated. The performance
analysis of MURUGA has also been carried out and compared with existing algorithms.
The results confirm that MURUGA performs better in terms of time taken to return fea-
sible paths.
Kim13 studied how to obtain Steiner trees appropriately for efficient multicast routing.
Kim first introduces a scheme for generating a new weighted multicast parameter by
efficiently combining two independent measures: cost and delay. The proposal method
was called weighted parameter for multicast trees (WPMT) algorithm. WPMT can be
adjusted by the weight, ω ∈ [0,1]. For instance, if ω approaches 0, then the delay of the
multicast tree may be relatively lower than the delay of other trees that are obtained as
ω approaches 1. Otherwise, as the weight approaches 1, the cost of the obtained tree may
be relatively lower compared with other trees. A case study also was given to illustrate
how to find an appropriate Steiner tree for each ω. The simulation results show that
the use of the proposed WPMT produces results similar to the k-minimum Steiner tree
algorithm.
Szymanski19 discussed a multicommodity maximum-flow–minimum-cost routing
algorithm. The algorithm computes maximum-flow routings for all smooth unicast traf-
fic demands within the capacity region of a network subject to routing cost constraints.
The edge cost can be a distance, reliability, congestion, or an energy metric. It is shown
that every network has a finite bandwidth–cost capacity. The bandwidth–distance and
the bandwidth–energy capacities are explored. The routing algorithm requires the for-
mulation of two linear programs (LPs). The first LP finds a multicommodity maximum
flow, when the flows are constrained to a subgraph of the network to enforce cost con-
straints. The second LP minimizes the routing cost, given that the maximum flow is
fixed. A related constrained multicast max-flow–min-cost algorithm is also presented, to
maximize the throughput of a multicast tree using network coding, subject to routing
cost constraints. These algorithms have polynomial-time solutions, whereas, traditional
216 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

multipath routing algorithms can be NP-Hard. The addition of routing cost constraints
can significantly reduce the size of the LPs resulting in faster solutions, with lower edge
utilizations and with higher energy efficiencies. The application of these algorithms to
route-aggregated video streams from cloud data centers in a future-Internet network, with
improved throughput, energy efficiency, and QoS guarantees is presented.
Sun21 presented a modified quantum-behaved particle swarm optimization (QPSO)
method for QoS multicast routing. Sun discussed that QoS multicast routing is converted
into an integer programming problem with QoS constraints and is solved by the QPSO
algorithm combined with loop deletion operation. The QPSO-based routing method,
along with the routing algorithms based on particle swarm optimization (PSO) and GA, is
tested on randomly generated network topologies for the purpose of performance evalua-
tion. The simulation results show the efficiency of the proposed method on the QoS rout-
ing problem and its superiority to the methods based on PSO and GA.

8.7 Conclusion
In this chapter, in order to give a more clear and practical reference about how the QoS
function is realized in ATN, we combined the ATN environment with IPv6, MIPv6, and
NEMO.
Starting from some basic knowledge, we first gave the requirement of QoS in ATN/IPS,
then we discussed QoS in two sections, since L2-QoS and L3-QoS are different but are
closely related. In Section 8.3, we gave an introduction of the next generation L-DACS1 sys-
tem, and mainly spoke about some RRM algorithm, and gave some figures to give a clear
description. This RRM algorithm is not referred in L-DACS1 specification but is very use-
ful to design a real system. L3-QoS and its application in ATN is introduced in Section 8.4
and as QoS is not defined in mobile environment, some useful research achievements is
given in Section 8.4.6.
QoS optimization is highly dependent on the mathematical graph model, hence, we have
discussed the unicast and multicast models in Sections 8.5 and 8.6, and we have also intro-
duced some recently published methods for further study. In fact, QoS should be consid-
ered not only based on mathematical models but also on practical network structure; how
to combine these two in ATN/IPS can be studied in the next step of research.

References
1. ICAO. Manual of technical provisions for the aeronautical telecommunication network (ATN),
9705/AN-956, Second Edition (Effective December 10, 1999), 2002.
2. Eurocontrol/FAA. Action plan 17 future communications study final conclusions and recom-
mendations report, Version 1.1, November 2007.
3. Eurocontrol/FAA. Communications operating concepts and requirements for the future radio
system, Eurocontrol/FAA, Technical report, May 2007.
4. Sajatovic, M. et  al. L-DACS1 system definition proposal: Deliverable D2. Brussels, Belgium:
Eurocontrol, Version 1, 2009.
Analysis and Modeling of the QoS Mechanism in ATN 217

5. Sajatovic, M. et al. Updated L-DACS1 system specification, SESAR P15 2, 2011. http://www.
ldacs.com/publications-and-links/.
6. Ayaz, S. Mobility and radio resource management in future aeronautical mobile networks. PhD
thesis. Erlangen, Germany: University Erlangen-Nürnberg, 2012.
7. Shreedhar, M. and Varghese, G. Efficient fair queueing using deficit round robin. ACM
SIGCOMM Computer Communication Review, 25(4), 231–242, 1995.
8. Lee, C.-W., Chen, M.C., and Sun, Y.S. Protocol and architecture supports for network mobility
with QoS-handover for high-velocity vehicles. Wireless Networks, 19(5), 811–830, 2013.
9. Kan, Z., Zhang, D., Zhang, R., and Ma, J. QoS in mobile IPv6. In Proceedings of International
Conferences on Info-Tech and Info-Net (ICII 2001), Beijing, China, 2001, Vol. 2, pp. 492–497.
10. Seite, P. et al. Quality-of-service option for proxy mobile IPv6. Internet Engineering Task Force.
(2014). https://tools.ietf.org/html/rfc7222.
11. Chen, S. and Nahrsted, K., An overview of quality of service routing for next-generation high-
speed networks: Problems and solutions. IEEE Network, 12(6), 64–79, November/December
1998.
12. Curado, M. and Monteiro, E. A survey of QoS routing algorithms. In Proceedings of the
International Conference on Information Technology (ICIT 2004), Istanbul, Turkey, 2004.
13. Kim, M., Choo, H., Mutka, M.W., Lim, H.J., and Park, K. On QoS multicast routing algorithms
using k-minimum Steiner trees. Information Sciences, 238, 190–204, 2013.
14. Medhi, D. Network Routing: Algorithms, Protocols, and Architectures. Boston, MA: Morgan
Kaufmann, 2010.
15. ICAO. Aeronautical Communications Panel (Working Group I), Manual for the ATN Using IPS
Standards and Protocols, Doc 9896, First Edition, April 2014. http://www.icao.int/safety/acp/
Pages/.
16. Brandes, S., Epple, U., Gligorevic, S., Schnell, M., Haindl, B., and Sajatovic, M. Physical
layer specification of the L-band digital aeronautical communications system (L-DACS1). In
Integrated Communications, Navigation and Surveillance Conference, Arlington, VA, 2009, p. 1.
17. Yin, P.Y., Chang, R.I., Chao, C.C., and Chu, Y.T. Niched ant colony optimization with colony
guides for QoS multicast routing. Journal of Network and Computer Applications, 40, 61–72, 2014.
18. Lu, T. and Zhu, J. Genetic algorithm for energy-efficient QoS multicast routing. IEEE
Communications Letters, 17(1), 31–34, 2013.
19. Szymanski, T.H. Max-flow min-cost routing in a future-internet with improved QoS guaran-
tees. IEEE Transactions on Communications, 61(4), 1485–1497, 2013.
20. Leela, R., Thanulekshmi, N., and Selvakumar, S. Multi-constraint QoS unicast routing using
genetic algorithm (MURUGA). Applied Soft Computing, 11(2), 1753–1761, March 2011.
21. Sun, J., Fang, W., Wu, X., Xie, Z., and Xu, W. QoS multicast routing using a quantum-behaved
particle swarm optimization algorithm. Engineering Applications of Artificial Intelligence, 24(1),
123–131, 2011.
9
Time Division Multiplexing in Satellite
Aeronautical Communications System

Sarhan M. Musa and Zhijun Wu

CONTENTS
9.1 Introduction......................................................................................................................... 219
9.2 Multiplexing Models and Results.................................................................................... 224
9.2.1 Synchronous TDM..................................................................................................225
9.2.2 Statistical TDM........................................................................................................ 227
9.3 Conclusion........................................................................................................................... 235
References...................................................................................................................................... 235

9.1 Introduction
Aeronautical communications has advanced different solution systems that allow aircraft
to maintain a link with the ground while in flight due to the growth of worldwide air
traffic and the increased need for communication safety. Today, we observe that satellite
solution systems are effective in supporting aeronautical communications. Indeed, satel-
lite communications is currently playing a major role toward the implementation of a
global communication infrastructure, especially given the explosive growth of wireless
technology. This chapter presents the analysis and simulation of blocking and clipping
probabilities for time division multiplexing (TDM) in a satellite aeronautical communica-
tions system. Specifically, we illustrate the evaluation of multiplexing systems in which
the number of input sources is greater than the number of available channels. For the case
of the blocking situation in synchronous TDM, we investigate the blocking probability
and the average number of busy channels that can be delivered. For the case of clip-
ping in statistical TDM, we examine the clipping probability and the expected number of
busy channels that can be delivered. We compare the blocking and clipping probabilities
for a fixed number of sources and different numbers of channels. We also compare the
expected number of busy available channels for synchronous TDM and the average
number of busy channels for statistical TDM methods for a fixed number of sources and
different numbers of channels.
Satellite communications was first deployed in the 1960s for military applications.
Satellites have played an important role in both domestic and international communi-
cations networks since the launching of the first commercial communication satellite by
NASA in 1965. They have brought voice, video, and data communications to areas of the
world that are not accessible with terrestrial lines. By extending communications to the
remotest parts of the world, virtually everyone can be part of the global economy.

219
220 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

An aircraft can be connected to the ground via a variety of different wireless access
technologies. A list of these technologies and their characteristics is provided in [1].
Satellite communications is not a replacement of existing terrestrial systems but rather
an extension of wireless systems. However, satellite communications has the following
merits over terrestrial communications [2]:

1.
Coverage: Satellites can cover a much larger geographical area than the traditional
ground-based system. They have a unique ability to cover the globe.
2.
High bandwidth: A Ka-band (27–40 GHz) can deliver throughput of gigabits per
second rate.
3.
Low cost: A satellite communications system is relatively inexpensive because
there are no cable-laying costs and one satellite covers a large area.
4.
Wireless communication: Users can enjoy untethered mobile communication any-
where within the satellite coverage area.
5.
Simple topology: Satellite networks have simpler topology, which results in more
manageable network performance.
6.
Broadcast/multicast: Satellites are naturally attractive for broadcast/multicast
applications.
7.
Maintenance: A typical satellite is designed to be unattended, requiring only mini-
mal attention by customer personnel.
8.
Immunity: A satellite system will not suffer from disasters such as floods, hur-
ricanes, fire, and earthquakes and will therefore be available as an emergency
service should terrestrial services be knocked out.

Based on the location of the orbit, satellites are divided into three categories: geostation-
ary Earth orbit (GEO), medium Earth orbit (MEO), and low Earth orbit (LEO) as shown in
Figure 9.1. A comparison of the three satellite types is given in Table 9.1.
In order to facilitate satellite communications and eliminate interference between dif-
ferent systems, international organizations govern the use of satellite frequency. The
International Telecommunication Union is responsible for allocating frequencies to satel-
lite services.

Earth

GEO
MEO
LEO

FIGURE 9.1
The three common types of satellites: geostationary Earth orbit, medium Earth orbit, and low Earth orbit.
Time Division Multiplexing in Satellite Aeronautical Communications System 221

TABLE 9.1
Comparison of Geostationary Earth Orbit, Medium Earth Orbit, and Low Earth Orbit
Type (km) Altitude Coverage Advantages Disadvantages
LEO 300–1,000 Spot Low path loss Less coverage
High data rate Need many satellites
Low delay Short orbital life
Low launch cost High Doppler
Less fuel Highly complex
MEO 1,000–10,000 Region Moderate path loss Multiple satellites
Moderate launch cost Moderate coverage
Less fuel Highly complex
GEO 36,000 Earth Global coverage High path loss
Need few satellites Long delay
Long orbital life Low data rate
Low Doppler High launch cost
Less complex Fuel for station keeping

TABLE 9.2
Satellite Frequency Allocations
Frequency Band Range (GHz)
L 1–2
S 2–4
C 4–8
X 8–12
Ku 12–18
K 18–27
Ka 27–40

The frequency spectrum allocations for satellite services are given in Table 9.2. In fact,
the signals between the satellite and the Earth stations travel along line-of-sight paths and
experience free-space loss that increases as the square of the distance.
Link budget of satellite is calculated for both the uplink from the gateway to the satel-
lite and the downlink from the satellite to the user terminal as shown in Figure 9.2. Link
budget of a satellite can be computed by the single link equation as

C ⎛ 1⎞⎛G ⎞⎛ 1 ⎞
= ( PTGT ) ⎜ ⎟ ⎜ R ⎟ ⎜ ⎟ (9.1)
N0 ⎝ L⎠ ⎝ T ⎠ ⎝ K ⎠

where
C/N0 is the signal to noise power ratio in dBHz
PTGT is the transmitter equivalent isotropic radiated power in dBW
PT is the carrier power
GT is the transmit antenna gain
1/L is the cumulative path loss
GR is the antenna gain of the receiver
T is the receiver system temperature
K is the Boltzmann constant
222 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Propagation losses

Extraneous noise
sources

Atmospheric losses

Uplink
Downlink

Earth station Earth station

FIGURE 9.2
(See color insert.) Satellite link budget.

A satellite band is divided into a number of separated portions: one for Earth-to-space
links (the uplink) and one for space-to-Earth links (the downlink). Separate frequencies are
assigned for the uplink and the downlink. Table 9.3 provides the general frequency assign-
ments for uplink and downlink satellite frequencies. We notice from the table that the
uplink frequency bands are slightly higher than the corresponding downlink frequency
band. This is to take advantage of the fact that it is easier to generate RF power within a
ground station than it is on board a satellite. In order to direct the uplink transmission to

TABLE 9.3
Typical Uplink and Downlink Satellite Frequencies
Uplink Frequencies (GHz) Downlink Frequencies (GHz)
5.925–6.425 3.700–4.200
7.900–8.400 7.250–7.750
14.00–14.50 11.70–12.20
27.50–30.0 15.70–20.20
Time Division Multiplexing in Satellite Aeronautical Communications System 223

a specific satellite, the uplink radio beams are highly focused. In the same way, the down-
link transmission is focused on a particular footprint or area of coverage.
Multiplexing is the concept of allowing the simultaneous transmission of multiple sig-
nals across a single data link. Frequency division multiplexing (FDM), wavelength divi-
sion multiplexing (WDM), and TDM are multiplexing techniques.
FDM is an analog technique that can be applied when the bandwidth of a link (in hertz)
is greater than the combined bandwidths of the signals to be transmitted. In FDM, signals
generated by each sending device modulate different carrier frequencies. These modulated
signals are then combined into a single composite signal that can be transported by a link.
WDM is a light technique that can be designed to use the high-data-rate capability of
fiber-optic cable. The optical fiber data rate is higher than the data rate of metallic trans-
mission cable, but using a fiber-optic cable for a single line wastes the available bandwidth.
Multiplexing allows us to combine several lines into one.
TDM is a digital process that allows several connections to share the high bandwidth of
a link. Instead of sharing a portion of the bandwidth as in FDM, time is shared. Each con-
nection occupies a portion of time in the link.
Multiple access technologies allow different users to utilize the satellite’s resources
of power and bandwidth without interfering with each other. Satellite communications
systems use different types of multiple access technology, including frequency division
multiple access (FDMA), time division multiple access (TDMA), and code division mul-
tiple access (CDMA). The access technology can vary between the uplink and downlink
channels.
The ability of multiple Earth stations or users to access the same channel is known as
FDMA. In FDMA, each user signal is assigned a specific frequency channel. One disad-
vantage of FDMA is that once a frequency is assigned to a user, the frequency cannot be
adjusted easily or rapidly to other users when it is idle. The potential for interference from
adjacent channels is another major shortcoming.
In TDMA, each user signal is allotted a time slot. A time slot is allocated for each peri-
odic transmission from the sender to a receiver. The entire bandwidth (frequency) is avail-
able during the time slot. This access scheme provides priority to users with more traffic
to transmit by assigning those users more time slots than it assigns to low-priority users.
Satellite providers will extend the capability and will employ multiple frequency TDMA.
If there are N frequencies, each offering M Mbps of bandwidth, then the total available
bandwidth during a time slot is N × M Mbps. Although FDMA techniques are more com-
monly employed in satellite communications systems, TDMA techniques are more com-
plex and are increasingly becoming the de facto standard.
In CDMA, users occupy the same bandwidth but use spread spectrum signals with
orthogonal signaling codes. This technique increases the channel bandwidth of the sig-
nal and makes it less vulnerable to interference. CDMA operates in three modes: direct
sequence, frequency hopping, and time hopping.
In recent years, significant effort has been made toward evaluating blocking probabili-
ties experienced by customers contending for a commonly shared resource. By definition,
the blocking probability is the probability that a connection service request is denied due
to insufficient network resources. Proportional differentiation models have been proposed
as effective methods for scalable differentiation service provision into optical WDM net-
works with blocking probability to various traffic classes [3–5].
Once a channel is assigned to a given talkspurt, the channel is held till the spurt ends.
The occurrence of freeze-out typically causes the initial part of a talkspurt to be clipped.
If a talkspurt sees no channel available upon its arrival, the initial portion of the talkspurt
224 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

will be clipped until a newly freed channel can be assigned. Clipping probability is used
in measuring video quality [6–8].
In this work, we will illustrate the multiplexing models used in satellite communications
systems, especially TDM, focusing on synchronous TDM and statistical TDM methods.

9.2  Multiplexing Models and Results

The technology of satellite aeronautical communications can support fixed and wireless
data, voice, and video communications, Internet connections, and enterprise networking.
Satellite communications successfully use a continuously transmitted signal of TDM for
the outbound (downlink) to improve of transmission quality instead of orthogonal FDM,
because of the linearity requirement on the power amplifier. TDMs are used in satellite
networks for maximum transmission capacity of a high-bandwidth line. Multiplexing
allows many communication sources to transmit data over a single physical line.
There are a number of multiplexing methods used in satellite communications systems.
One of the commonly used methods in such systems is the TDM technology as in [9–12].
Blocking and statistical probabilities are applied for TDM multiplexing as in [13–22]. In
this chapter, we will analyze and simulate the blocking and statistical probabilities of
TDM applied to satellite communications systems.
Figure 9.3 shows a satellite communications system using TDM technology in interac-
tive and sending/receiving different applications based on the importance of response
time. TDM is used in the outbound link between the source (sender or host) and the user
(receiver). A TDM system is a high-speed data stream scheme that acts at layer 1 (physi-
cal layer) of the Open Systems Interconnection model and at layer 4 (network interface)
of the Transmission Control Protocol over Internet Protocol model. In TDM technology,
users take turns in a predefined way, each one periodically getting the entire bandwidth

Satellite

Aircraft

Inbound

Outbound TDM

Cluster
VSAT
control
&
process
User

Comm.
Hub
control
&
process
Host

FIGURE 9.3
Satellite communications system with TDM.
Time Division Multiplexing in Satellite Aeronautical Communications System 225

Frame i
1 1
. . .
2 2
Ch 1 Ch 2 Ch c

s Multiplexer Demultiplex s

FIGURE 9.4
A time division multiplexer with s inputs and c channels.

for a portion of the total scanning time. The input source s is divided into frames, and
each frame is subdivided into time slots (channels), c, where each channel is allocated
to one input as shown in Figure 9.4. Packets arrive on s lines, and the multiplexer scans
them, forming a frame with c channels on its outgoing link. There are two different
types of TDMs to deal with the different ways in which channels of frames could be
allocated as synchronous and statistical (asynchronous) and it has been used with other
techniques as solution for satellite communications networks such as TDMA, FDMA,
and PAMA.

9.2.1  Synchronous TDM

In synchronous TDM, a frame is divided in fixed-sized channels and channels are allo-
cated to input sources in a fixed way. The Quality of Service (QoS) of synchronous
TDM is based on how its transmission system is set up. For example, the multiplexer
is inefficient when the number of users is greater than the available channels. This is
true since the multiplexer scans all input source lines without exceptions and the scan-
ning time for each input source line (each connected to a user) is reallocated, as well
as this time for a particular input source line is not altered by the system control. The
scanner should stay on that input source line, whether or not there are data for scanning
within that time slot. A synchronous TDM can also be programmed to produce same-
sized frames; the lack of data in any channel potentially creates changes to average bit
rate on the ongoing link.
To analyze a synchronous multiplexer, let ta and td be the mean time for active input
source and the mean time for idle input source, respectively. Let us assume that the values
of ta and td are random and exponentially distributed. (This assumption is based on expe-
rience.) Also, consider a TDM with a number of requesting input sources, s, greater than
available channels, c, where s > c; the TDM will react by blocking. The unassigned input
sources are not transmitted and therefore remain inactive. The probability that an input
source is active, ρ, can be obtained by ρ = ta/(ta + td).
Let Ps(j) be the probability of j different inputs out of s that are active:

⎛ s⎞
Ps ( j) = ⎜ ⎟ ρ j (1 − ρ ) , (9.2)
s− j

⎝ j⎠

∑
c
where 1 ≤ j ≤ s. We know that ps ( j) can never be equal to 1, and in fact we must have
j=0

∑
s
ps ( j) = 1. This can lead to normalization of Ps(j) over c available channels. Thus, the
j=0
probability of j different output of c available channels, Pc(j), is
226 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

j
⎛ s⎞ ⎛ ρ ⎞
⎜⎝ j ⎟⎠ ⎜⎝ 1 − ρ ⎟⎠
Pc ( j) = i
, (9.3)
⎛ s⎞ ⎛ ρ ⎞
∑
c

⎝ j ⎟⎠ ⎜⎝ 1 − ρ ⎟⎠
i=0 ⎜

where 0 ≤ j ≤ c and 0 ≤ i ≤ c. The blocking probability Ps(c) can be obtained when j = c:

c
⎛ s ⎞ ⎛ ta ⎞
⎜⎝ c ⎟⎠ ⎜⎝ t ⎟⎠
d
Ps (c) = i
, (9.4)
∑ ⎛ s ⎞ ⎛ ta ⎞
c
⎜ ⎟⎜ ⎟
i=0 ⎝ i ⎠ ⎝ td ⎠

where in general 0 ≤ i ≤ c.
Figure 9.5 shows the blocking probability for the fixed number of sources (s = 10) and
different numbers of channels (c = 1, c = 4, and c = 7), whereas Figure 9.6 shows the block-
ing probability for the fixed number of sources (s = 10) and different numbers of channels
(c = 2, c = 5, and c = 8). The blocking probability clearly rises with the increased utiliza-
tion, ρ, for all three cases; and also it is higher when a fewer number of channels, c, are
available.
Then, we can calculate the expected number of busy channels for the multiplexer,
Ec(b), by

Ps (c)
1

0.8
c=1

0.6 c=4

c=7
0.4

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.5
Comparison of blocking probability, Ps(c), versus probability of active input source, ρ, with 0 ≤ ρ ≤ 1, for s = 10,
c = 1, 4, and 7.
Time Division Multiplexing in Satellite Aeronautical Communications System 227

Pi (l)
1

0.8 c=2

0.6

c=5
0.4 c=8

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.6
Comparison of blocking probability, Ps(c), versus probability of active input source, ρ, with 0 ≤ ρ ≤ 1, for s = 10,
c = 2, 5, and 8.

j
⎛ s⎞ ⎛ t ⎞
∑
c
j⎜ ⎟ ⎜ a ⎟
j=1 ⎝ j ⎠ ⎝ td ⎠
Ec (b) = i
, (9.5)
∑ ⎛ s ⎞ ⎛ ta ⎞
c
⎜ ⎟⎜ ⎟
i=0 ⎝ i ⎠ ⎝ td ⎠

⎛ ρ ⎞ ⎛ ta ⎞
where 1 ≤ j ≤ c, 0 ≤ i ≤ c, and ⎜ = .
⎝ 1 − ρ ⎟⎠ ⎜⎝ td ⎟⎠

Figure 9.7 shows the expected number of busy (unavailable) channels for fixed number of
sources (s = 10) with different numbers of channels (c = 1, c = 4, and c = 7), while Figure 9.8
shows the expected number of busy (unavailable) channels for fixed number of sources
(s = 10) with different numbers of channels (c = 2, c = 5, and c = 8). The expected number of
busy channels varies in its maximum values based on the interval of utilization.

9.2.2  Statistical TDM

Statistical TDM has high efficiency because a frame’s time slots are dynamically allocated,
based on demand, and it removes all the empty slots on a frame. But it is difficult to give
a guarantee QoS, because of the requirement that additional overhead be attached to each
outgoing channel. These additional data are needed because each channel must carry infor-
mation about which input source line it belongs to. The frame length is available not only
because of different channel sizes but also because of the possible absence of some channels.
We consider ta and td as random and exponentially distributed. Also, consider a TDM
with a number of requesting input sources, s, greater than available channels, c, where s >
c; the TDM will react by clipping; the unassigned input sources are partially transmitted. If
228 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Ec (b)

10
c=7

c=4
6

2
c=1

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.7
Comparison on expected numbers of busy available channels, Ec(b), versus probability of active input source, ρ,
where 0 ≤ ρ ≤ 1, for s = 10, c = 1, 4, and 7.

Ec (b)

12
c=5
10

6 c=8

c=2
2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.8
Comparison on expected numbers of busy available channels, Ec(b), versus probability of active input source, ρ,
where 0 ≤ ρ ≤ 1, for s = 10, c = 2, 5, and 8.
Time Division Multiplexing in Satellite Aeronautical Communications System 229

more than c input channels are active, we can dynamically choose c out of s active sources
and temporarily block other sources. In this temporary blocking, the source is forced to
clip or lose data for a short period of time, where the amount of data lost depends on ta, td,
s, and c, but the source may return to a scanning scenario if a channel becomes free.
The clipping probability, Pi(l), or the probability that an idle source finds at least c chan-
nels busy at the time it becomes active can be calculated by considering all s sources minus
1 (the examining source):
s−1

Pi (l) = ∑ ⎛⎜⎝ s −i 1 ⎞⎟⎠ ρ (1 − ρ)

i s−1−i
, (9.6)
i=c

where c ≤ i ≤ s − 1.
Figure 9.9 shows the clipping probability for a fixed number of sources (s = 10) and dif-
ferent numbers of channels (c = 1, c = 4, and c = 7), while Figure 9.10 shows the clipping
probability for a fixed number of sources (s = 10) and different numbers of channels (c = 2,
c = 5, and c = 8). The clipping probability of two channels has the highest clipping prob-
ability compared to five and eight channels.
Clearly, the average number of used channels, Ac(u), is
j
⎛ s⎞ ⎛ t ⎞
∑
c
j ⎟⎜ a⎟
j=1 ⎜
⎝ j⎠⎝ t ⎠ d
Ac (u) = i
, (9.7)
⎛ s ⎞ ⎛ ta ⎞
∑
c

⎝ j ⎟⎠ ⎜⎝ t ⎟⎠
i=0 ⎜ d

Pi (l)
1

0.8
c=1

c=4
0.6

c=7
0.4

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.9
Comparison of clipping probability, Pi(l), versus probability of active input source, ρ, where 0 ≤ ρ ≤ 1, for s = 10,
c = 1, 4, and 7.
230 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Pi (l)
1

c=2
0.8

c=5
0.6

c=8

0.4

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.10
Comparison of clipping probability, Pi(l), versus probability of active input source, ρ, where 0 ≤ ρ ≤ 1, for s = 10,
c = 2, 5, and 8.

Ac (u)
8
c=8

c=5
5

c=2
2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.11
Comparison of average number of used channels, Ac(u), versus probability of active input source, ρ, where
0 ≤ ρ ≤ 1, for s = 10, c = 2, 5, and 8.
Time Division Multiplexing in Satellite Aeronautical Communications System 231

where 0 ≤ i ≤ c and 1 ≤ j ≤ c. Figure 9.11 shows the average number of used channels for a
fixed number of sources (s = 10) and different numbers of channels (c = 2, c = 5, and c = 8).
The average number of used channels of eight channels has the highest average number of
used channels compared to the ones for two and five channels.
The average number of busy channels, Ac(b), is
j
⎛ s⎞ ⎛ t ⎞
∑
c
j ⎟⎜ a⎟
j=1 ⎜
s
⎝ j⎠⎝ t ⎠ ⎛ ⎞
Ac (b) =
⎛ s ⎞ ⎛ ta ⎞
d
i
+c ∑ ⎜⎝ sj ⎟⎠ ρ (1 − ρ)
j s− j
, (9.8)
∑
c
j=c+1
⎝ j ⎟⎠ ⎜⎝ t ⎟⎠
i=0 ⎜ d

where 0 ≤ i ≤ c, 1 ≤ j ≤ c, and c + 1 ≤ j ≤ s.
Figure 9.12 shows the average number of busy channels for a fixed number of sources
(s = 10) and different numbers of channels (c = 2, c = 5, and c = 8). The average number of
busy channels for all cases is almost the same up to ρ = 0.25, but it differs for ρ > 0.25.
Figures 9.13 through 9.15 show the comparison between blocking and clipping
probabilities for a fixed number of sources (s = 10) and different numbers of channels (c = 2,
c = 5, and c = 8, respectively). We observe that the blocking probability is greater than the
clipping probability for 2 and 5, but it varies at channel 8.
Figures 9.16 through 9.18 show the expected number of busy available channel for syn-
chronous TDM and the average number of busy channel for statistical TDM methods for a
fixed number of sources (s = 10) and different numbers of channels (c = 2, c = 5, and c = 8,
respectively). We observe that the average number of channels and the expected number
of busy available channels vary in the different utilization numbers.

Ac (b)

c=8
14

12

c=5
10

c=2
4

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.12
Comparison of average number of busy channels, Ac(b), versus probability of active input source, ρ, where 0 ≤
ρ ≤ 1, for s = 10, c = 2, 5, and 8.
232 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Pi(l) or Ps(c)
1

Pi(l)

0.8 Ps(c)

0.6

0.4

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.13
Comparison of blocking probability, Ps(c), and clipping probability, Pi(l), versus probability of active input
source, ρ, with 0 ≤ ρ ≤ 1, for s = 10, c = 2.

Pi(l) or Ps(c)
1

0.8 Ps(c)

Pi(l)

0.6

0.4

0.2

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.14
Comparison of blocking probability, Ps(c), and clipping probability, Pi(l), versus probability of active input
source, ρ, with 0 ≤ ρ ≤ 1, for s = 10, c = 5.
Time Division Multiplexing in Satellite Aeronautical Communications System 233

Pi(l) or Ps(c)
1

0.8

0.6

0.4

Ps(c)
0.2
Pi(l)

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.15
Comparison of blocking probability, Ps(c), and clipping probability, Pi(l), versus probability of active input
source, ρ, with 0 ≤ ρ ≤ 1, for s = 10, c = 8.

Ac(b) or Ec(b)
4

Ac(b)

Ec(b)

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.16
Comparison of expected number of busy available channels, Ec(b), and average number of busy channels, Ac(b),
versus probability of active input source, ρ, where 0 ≤ ρ ≤ 1, for s = 10, c = 2.
234 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Ac(b) or Ec(b)

12
Ec(b)
Ac(b)
10

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.17
Comparison of expected number of busy available channels, Ec(b), and average number of busy channels, Ac(b),
versus probability of active input source, ρ, where 0 ≤ ρ ≤ 1, for s = 10, c = 5.

Ac(b) or Ec(b)

14

12 Ac(b)

10

6
Ec(b)

0 ρ
0 0.2 0.4 0.6 0.8 1

FIGURE 9.18
Comparison of expected number of busy available channels, Ec(b), and average number of busy channels, Ac(b),
versus probability of active input source, ρ, where 0 ≤ ρ ≤ 1, for s = 10, c = 8.
Time Division Multiplexing in Satellite Aeronautical Communications System 235

9.3 Conclusion
We presented the analysis of TDM applied to satellite aeronautical communications sys-
tems when input sources are greater than available channels. The analysis of blocking
and clipping probabilities for TDMs was successfully achieved, and results of the analy-
sis were generated. For blocking cases in synchronous TDMs, we illustrated the block-
ing probability and the average number of busy channels that could be delivered. For
the clipping in statistical TDM, we examined the clipping probability and the expected
number of busy channels. We compared the blocking and clipping probabilities for
a fixed number of sources and different numbers of channels. We also compared the
expected number of busy available channel for synchronous TDM and the average num-
ber of busy channels for statistical TDM methods for a fixed number of sources and dif-
ferent numbers of channels.

References
1. Airlines Electronic Engineering Committee (AEEC). ARINC Report 811: Commercial aircraft
information security concepts of operation and process framework, Aeronautical Radio Inc.,
p. 138, December 2005.
2. M. N. O. Sadiku, Optical and Wireless Communications: Next Generation Networks. Boca Raton, FL:
CRC Press, pp. 185–217, 2002.
3. Y. Chen, M. Hamdi, and D. H. K. Tsang, Proportional QoS over WDM networks: Blocking
probability, Proceedings of Sixth IEEE Symposium on Computers and Communications, pp. 210–215,
Hammamet, July 2001.
4. A. Bianco, G. Galante, and M. Mellia, Analysis of call blocking probability in TDM/WDM net-
works with transparency constraint, IEEE Communications Letters, 4(3), 104–106, March 2000.
5. Y. C. Huei and P. H. Keng, A shared time-slot router architecture for TDM wavelength optical
WDM, Proceedings of IEEE Seventh Malaysia International Conference on Communication, Vol. 2,
pp. 672–677, Malaysia, November 2005.
6. C. J. Weinstein, Fractional speech loss and talker activity model for TASI and for packet-
switched speech, IEEE Transactions on Communications, 26, 1253–1257, August 1978.
7. S. Q. Li, A new performance measurement for voice transmission in burst and packet switch-
ing, IEEE Transactions on Communications, 35(10), 1083–1094, October 1987.
8. A. Gatherer and M. Polley, Controlling clipping probability in DMT transmission, Proceedings of
the 31st Asilomar Conference on Signals, Systems and Computers, Vol. 1, pp. 578–584, Pacific Grove,
CA, November 1997.
9. O. R. Ramires, Use of TDM/PAMA satellite based communication as an option for the aeronau-
tical communication, IEEE National Aerospace and Electronics Conference, pp. 683–686, Dayton,
OH, October 2000.
10. Y. Chaehag, TDM framing for gap filler operation in satellite digital multimedia broadcasting
System A, IEEE 59th Vehicular Technology Conference, Vol. 5, pp. 2782–2786 May 2004.
11. K. Zhang and S. Hryckiewicz, An integrated approach for IP networking over the wideband
Gapfiller satellite, IEEE on Military Communications Conference (MILCOM), Vol. 3, pp. 1556–1561,
November 2004.
12. K. Eng, and A. Acampora, Fundamental conditions governing TDM switching assignments in
terrestrial and satellite networks, IEEE Transactions on Communications, 35(7), 755–761, July 1987.
236 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

13. W. K. Lai, Y. J. Jin, H. W. Chen, and C. Y. Pan, Channel assignment for initial and handoff
calls to improve the call-completion probability, IEEE Transactions on Vehicular Technology, 52(4),
876–890, July 2003.
14. J. W. Chong; B. C. Jung, and D. K. Sung, Statistical multiplexing-based hybrid FH-OFDMA
system for OFDM-based UWB indoor radio access networks, IEEE Transactions on Microwave
Theory and Techniques, 54(4), 1793–1801, June 2006.
15. A. I. Elwalid and D. Mitra, Statistical multiplexing with loss priorities in rate-based conges-
tion control of high-speed networks, IEEE Transactions on Communications, 42(11), 2989–3002,
November 1994.
16. M. Balakrishnan, R. Cohen, E. Fert, and G. Keesman, Benefits of statistical multiplexing in
multi-program broadcasting, International Broadcasting Convention, 1, 560–565, September 1997.
17. J. Liebeherr, S. Patek, and E. Yilmaz, Tradeoffs in designing networks with end-to-end statisti-
cal QoS guarantees, Eighth International Workshop on Quality of Service (IWQOS), pp. 221–230,
Pittsburgh, PA, June 2000.
18. Z. L. Zhang, J. Kurose, J. D. Salehi, and D. Towsley, Smoothing, statistical multiplexing, and
call admission control for stored video, IEEE Journal on Selected Areas in Communications, 15(6),
1148–1166, August 1997.
19. W. C. La and S. Q. Li, Statistical multiplexing and buffer sharing in multimedia high-speed net-
works: A frequency-domain perspective, IEEE/ACM Transactions on Networking, 5(3), 382–396,
June 1997.
20. A. Iera, A. Molinaro, S. Marano, and D. Mignolo, Statistical multiplexing of heterogeneous
traffic classes in ATM-satellite based networks, IEEE Wireless Communications and Networking
Conference (WCNC), Vol. 1, pp. 174–178, New Orleans, LA, September 1999.
21. L. Tancevski and I. Andonovic, Block multiplexing codes for incoherent asynchronous all-
optical CDMA using ladder network correlators, IEE Proceedings on Optoelectronics, 142(3),
125–131, June 1995.
22. M. Zafer and E. Modiano, Blocking probability and channel assignment in wireless networks,
IEEE Transactions on Wireless Communications, 5(4), 869–879, April 2006.
10
ATN Transmission Control Algorithm
Based on Service-Oriented Architecture

Haitao Zhang

CONTENTS
10.1 Features and Composition of ATN Network.................................................................. 238
10.1.1 Composition of ATN.............................................................................................. 238
10.1.2 Data Information and Related Systems of ATN................................................. 239
10.1.3 Structure of ATN Network System...................................................................... 240
10.1.4 ATN and SWIM....................................................................................................... 241
10.2 Introduction of SOA........................................................................................................... 241
10.2.1 Service-Oriented Architecture.............................................................................. 241
10.2.2 SOA Logical Layers................................................................................................ 242
10.2.3 ESB—Core Infrastructure of SOA........................................................................ 244
10.2.4 Industry Benefits from Implementing SOA........................................................ 245
10.3 Research to Improve the Stability of SWIM Based on SOA......................................... 246
10.3.1 General Introduction to Releasing the Thread Blocked of
Application Server.................................................................................................. 246
10.3.2 Adopting Aggressive Timer Settings: Methods Based on Self-Contained
Queue Model........................................................................................................... 247
10.3.2.1 Application of Aggressive Timer Settings............................................ 247
10.3.2.2 Implementation of Discretization of Timeout Probability
Attribute of Queue Model by KDD Language Field........................... 248
10.3.3 Optimizing the Deployment of SOA Services or Applications.......................254
10.3.3.1 Optimizing and Deploying Tightly Coupled SOA Services..............254
10.3.3.2 Advantages of Optimized Deployment................................................254
10.3.3.3 Optimized Deployment under High Workloads................................ 255
10.3.3.4 Preferable Service Deployment Solution Based on Clustering
Analysis of Log......................................................................................... 256
10.4 Cluster Analysis on the Hadoop Cloud Computing Platform..................................... 257
10.4.1 Introduction of Hadoop Cloud Computing Platform....................................... 257
10.4.2 Exploration of Hadoop Clusters and Internet.................................................... 257
10.4.2.1 Basic Principles of Operation of Hadoop Clusters.............................. 257
10.4.2.2 Structure of the Typical Hadoop Cluster.............................................. 258
10.4.2.3 Hadoop Workflows.................................................................................. 258
10.4.3 MapReduce Data Flow........................................................................................... 259
10.4.4 Design and Implementation of Parallel K-Means Clustering Algorithm
Based on Hadoop Platform................................................................................... 261
10.4.4.1 Function Map()......................................................................................... 261

237
238 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

10.4.4.2 Function Combine()................................................................................. 262

10.4.4.3 Function Reduce().................................................................................... 262
10.5 Conclusion........................................................................................................................... 263
References...................................................................................................................................... 265

10.1  Features and Composition of ATN Network

The Aeronautical Telecommunications Network (ATN) is an important part of the com-
munications, navigation, and surveillance/air traffic management system (CNS/ATM) and
the basis for a new generation of navigation systems. ATN adopts seven communication
models formulated by the International Organization for Standardization (ISO).

10.1.1  Composition of ATN

ATN consists of three subnets: ground–ground communication subnet, air–ground com-
munication subnet, and air–air communication subnet. Currently, the ground/air com-
munication subnet of ATN is mainly made up of very high frequency (VHF) data link,
aeronautical mobile satellite communications (AMSS), and secondary surveillance radar
mode S data link [1].

1.
VHF data link: The VHF system works in 118–136  MHz frequency band and its
communication coverage radius is inside 200 nautical miles. There are four modes,
mode 1, mode 2, mode 3, and mode 4 in the VHF system, and mode 2 and mode
4 have been recognized as an international standard by ICAO, which are also
very promising technologies. VHF mode 2 adopts D8PSK modulation mode; each
channel bandwidth is 25 kHz, the bit rate is 31.5  kbps, and p-persistent carrier
sense shares channel access whose probability p is 13/256. VHF mode 2 can sup-
port voice and data communications, which is now widely applied in air–ground
communication systems of domestic and international civil aviation.
2.
AMSS: As a powerful means of communication for aircraft flying over oceans,
deserts, Antarctic, Arctic, and other areas where low altitude communication
infrastructure is weak, AMSS can provide voice and data communications for
aircraft in areas where VHF and HF cannot cover. Currently, INMARSAT is able
to provide satellite communications services worldwide, and its satellite beam
can cover most of the world’s busiest routes, providing services for many users
through ground stations and ground earth stations (GESs) and providing com-
munication with channel rates of 9600, 64, and 256 kbps for the users.
3.
Secondary surveillance radar mode S: Secondary surveillance radar A or C mode
coding is limited in number and the exchangeable information (ID, height) is
little. All aircraft will receive inquiry signals and respond to them at the same
time, which results in mutual signal interference and loss. Mode S is compatible
with modes A/C, and the first 24 bits of each inquiry signal serves as an address
code of the aircraft, so the identifiable number of aircraft can reach 224, which is
more than 4000 times as the current A mode does, being enough to implement an
address for an airplane.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 239

10.1.2  Data Information and Related Systems of ATN

Information transmitted between the aircraft and the ground includes air traffic
management (ATM) information, airlines operation information, and personal communi-
cation information of travelers by air. According to the priority, the ATM information takes
precedence over airlines operation information and airlines operation information takes
precedence over personal communication information. The ATM information itself also
includes various types of messages. ATM information in air-to-ground communication
includes automatic dependent surveillance messages, navigational intelligence messages,
directory services messages, and the pilot–controller data link communication messages.
These messages also have different QoS requirements. ICAO divided air traffic manage-
ment information into eight types from A to H according to different transmission delay
requirements, where A type is reserved and B to H are applied to different air–ground
messages, as is shown in Table 10.1.

1.
Automatic dependent surveillance (ADS): Automatic dependent surveillance-
broadcast (ADS-B) is an important part of a new generation of air traffic man-
agement systems, as ADS-B, compared with conventional radar, provides higher
security, costs less, and is superior in operation. Aircraft equipped with ADS-B
can automatically keep sending the ADS information, including the aircraft’s
identification code, latitude, longitude, velocity, direction, altitude, time, and other
auxiliary information to the ground controller’s ADS terminal or other aircraft.
The current situation in the air can also be shown on a display terminal on the
ground and other aircraft ground controllers can promptly detect the introduc-
tion errors of waypoints and ATC loop error, and carry out compliance overseeing
for the current flight plan and deviation detection based on ADS information; this
will help find out the deviation of the release path of the aircraft flight path and
provide better continuity for the path and track position, velocity, and acceleration
estimation to prevent collision, which in turn ensures flight safety intervals and
reduces the number of false alarms.
2.
Controller–pilot data link communications (CPDLC): CPDLC enables pilots and con-
trollers to carry out two-way communication directly via a data link so that they can
send a variety of requests and control commands. These requests and commands

TABLE 10.1
Transmission Delay Requirements of ATM Information
One-Way Transmission Delay in ATN
(Success Rate of Transmission Is 95%,
and Unit Is Second) Types of ATM Information
Reserved A
4.5 B
7.2 C
13.5 D
18 E
27 F
50 G
100 H
240 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

are implemented as data messages, thus avoiding ambiguity caused by direct voice
communication. Using the data link can increase system capacity and reduce the
workload of controllers and pilots.
3.
Flight information system (FIS): This service supports pilots with all sorts of infor-
mation during flight. At present, information is transmitted through voice with
a small amount of information and a long time of transmission. There are two
ways of using the data link: broadcast and request type. The information would
be stored in the database after the aircraft receives it for pilots to check. FIS can
provide a variety of text or graphical information on weather, airport, and aircraft
equipment.

10.1.3  Structure of ATN Network System

The ATN network system is divided into end system (ES) and intermediate system (IS), as
shown in Figure 10.1. The end system contains each user’s computer unit in the ATN, and

ES
Management processor Data processor Input processor

Airborne subnet

Air-to-air
BIS
data link
Router
Other aircraft

Air-to-air Air-to-air Air-to-air

data link a data link b data link c

BIS
Router Router

Airlines
Service
ground ATS subnet
provider
subnet

ES
Aviation operation AOC Weather ATC Air intelligence Weather Radar
database database database

FIGURE 10.1
Structure of ATN.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 241

provides end-to-end communications services for upper applications, and also provides
user interface for man–machine interaction. Airborne ES includes display and control sys-
tems in aircraft cockpit and operator’s workbenches in aircraft task electronic systems
such as radar, command guidance, and other workbenches. Ground ES includes all types
of end users in ground communication subnets, such as various ground control towers
and ground stations of ATM systems. By ATN, each end system ES carries out point-to-
point communication with other end ESs in the network system. The intermediate system
is the router of ATN. The routing domain of ATN network is divided into intradomain
routing and interdomain routing, so an intermediate system is subdivided into Border
Intermediate System (BIS) and IS. BIS is used to transfer data in interdomain subnets and
IS is used to transfer data in intradomain subnets.

10.1.4  ATN and SWIM

System-wide information management (SWIM) is a core technology to share and exchange
information of air traffic control in air transportations around the present world. Next
Generation Air Traffic Management in the United States and Single European Sky ATM
Research (SESAR) adopted SWIM to build and implement a framework for information
exchange, which has proven the functionality of service-oriented architecture (SOA) to
SWIM [2].
ATN system architecture works for the global deployment of ATM applications and ser-
vices, while SWIM is a solution to technical defects of the ATN system architecture and the
original system, which is the way of information management integrating and covering
ATN from the network.
SWIM’s goal is to integrate information of the virtual ATM network, and its solution is
carried out system-wide rather than in a specific system or interface. SWIM, as a set of net-
work infrastructure technology for information sharing and exchange, is expected to meet
all kinds of aviation accuracy and timeliness of information interaction between users,
mainly based on information and data as the center, through a loosely coupled architec-
ture with strong openness, flexibility and robustness, that is, SOA.
Current ATN infrastructure designed based on the past communication mechanism
(e.g., X.25) reaches a limited bandwidth and few manufacturers are willing to produce it.
However, as SWIM adopts the standardized TCP/IP protocol, which has reached a higher
bandwidth, the information sharing and collaborative decision making can be realized
through a standardized interface. The Federal Aviation Administration (FAA) has begun
parallel deployment of ATN and SWIM and has planned to achieve full integration of
ATN and SWIM in the long term. At the same time, SESAR plans the completion of the
European network deployment of IP-based ground data exchange, which lays the founda-
tion for the full deployment of SWIM services, before 2025.

10.2  Introduction of SOA

10.2.1  Service-Oriented Architecture [3]
In one common model of application interaction, an information provider interacts with
an information requester, which is based on SOA, with an information discovery process
assisting.
242 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Discovery

se Fin
erti d
Adv
Common support services

Interact
Provider Requester

FIGURE 10.2
Service-oriented model of application interaction.

In SOA, the provider could be any service and the requester could be any consumer. The
interaction between them is an easily locatable service interface. The provider offers its
services through a registry where the requester can find information sources it needs and
find out the location of the provider that will interact with it. The services are located with
service descriptions, which is the key point of SOA (Figure 10.2).
When information providers are multiple, the requester can pick up one according to
the policy or other standards. This is a form of redundancy that could be used to improve
information availability.
There are no restrictions on the interaction between information provider and
requester, which includes direct or mediated, push or pull, point to point or point to
multipoint, etc.
In an ideal condition, information services descriptions are machine-readable, which
allows the discovery and access of the dynamic information by application requesting.

10.2.2  SOA Logical Layers

Sajjad [4] proposes a seven-layer composition of the SOA architecture. According to this
study, SOA was shown as a partially layered architecture that composes of services aligned
with business process. It could be shown in Figure 10.3.
In this SOA architecture, the large-grained enterprise components or business-line com-
ponents carry out the services, implementing the functionality and ensuring the working
of the services. Combining these services to composite applications can support business
process flows. In an integrated architecture, the routing, mediation, and transformation of
these services, components, and processes are to be supported by enterprise service bus
(ESB). The deployed services must be monitored and managed for quality of service and
nonfunctional requirement.
The SOA layered architecture is described layer by layer as follows:

Layer 1: Operating system layer. This is the basic layer, consisting of legacy sys-
tems that the programs that have been developed with an outdated technology
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 243

Layer 7: Quality of services (QoS)

Layer 5: Access or presentation layer

Layer 6: Integration (Enterprise

Service Bus—ESB)
Layer 4: Business process composition layer

Layer 3: Services layer

Layer 2: Enterprise components layer

Layer 1: Operational system

FIGURE 10.3
SOA logical layers.

make up for the vast majority of programs in many user application environ-
ments, and existing customer relationship management (CRM) and enterprise
resource planning (ERP) applications, older object-oriented applications, and
business intelligence applications. SOA composite-layered architecture can
integrate these applications with existing systems and service-based integration
techniques.
Layer 2: Enterprise components layer. This layer is responsible for realizing function-
ality and maintaining the QoS of the exposed services. This layer typically uses
application servers to implement the component, workload management, high
availability, and load balancing.
Layer 3: Services layer. The services needed for business applications are found in
this layer. They are composed of dynamic or static formats or a composition of
both. Therefore, the enterprise components can provide services at runtime using
the functionality provided by their interfaces. The interfaces can exist indepen-
dently as an exposed service description in this layer.
Layer 4: Business process composition layer. The services offered in layer 3 are defined
in this layer. They are bundled into a flow to work together as a single application.
These applications support special cases and business processes. IBM WebSphere
Business Integration Model or Websphere Application Developer Integration
Edition can be used for designing these application flows.
Layer 5: Access or presentation layer [5] explained this layer that gets more useful
though being not relevant to the architecture of SOA. Still some applications seek
to adopt Web Services at the application interface or presentation layer.
Layer 6: Integration (enterprise service bus [ESB]) layer. This layer provides a position
independent mechanism for integration. The layer integrates services by intro-
ducing a set of reliable performance convergence such as intelligence routing,
protocol mediation, and other transformation mechanisms. For instance, the Web
Services Description Language (WSDL) develops binding containing addresses
offering services.
Layer 7: Quality of Service (QoS). This layer is used to monitor, manage, and maintain
the security, performance, and availability, which is mainly a background process
with sense-and-response mechanisms and monitor tools for SOA state, including
the important standards’ implementation of Web Service Management (WSM).
244 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

10.2.3  ESB—Core Infrastructure of SOA

ESB is the combination of middleware technology with Web Services and the core infra-
structure of SOA. The concept has been developed in analogy to the bus concept found
in computer hardware architecture combined with the modular and concurrent design
of high-performance computer operating systems. The motivation was to find a stan-
dard, structured, and general purpose concept for describing implementation of loosely
coupled software components (called services) that are expected to be independently
deployed, running, heterogeneous, and disparate within a network. ESB is also the
intrinsically adopted network design of the World Wide Web and the common imple-
mentation pattern for SOA [6]. ESB, as a service intermediary, forms a chain of the ser-
vice users → ESB service proxy → the service provider. The intermediary functions in
various ways in different applications:

1.
Decoupling intermediary: Customer neither knows nor cares about the actual
identity of service provider, physical location, and transmission protocol and
interface definition. Interactive integration code extracted outside the busi-
ness logic, the central ESB platform makes central declarative definitions. ESB
platform carries out protocol conversion (Web Services, http, JMS, etc.), mes-
sage conversion (conversion, enrichment, and filtration), and message routing
(synchronous/asynchronous, publish/subscribe, content-based routing, branch-
ing and aggregation, etc.).
2.
Service intermediary: ESB platform provides basic services in services interaction
as an intermediary. ESB platform implements SLA (reliability assurance, load bal-
ancing, flow control, cache, business control, and encrypted transmission), service
management monitor (exception handling, service call and message data records,
condition monitor of systems and services, and ESB configuration management),
and unified security management (as it is hard to be achieved).
3.
Service orchestration: Several services are orchestrated to form a new service. ESB
supports a visualized form to define a new composite service process (workflow,
BPEL, or code-level orchestration).

Therefore, the basic functions of ESB are data transmission, message protocol conversion,
and routing. These three core functions are always provided by ESB in the integration of
heterogeneous systems. Although SOA can also be carried out without ESB by means of
SCA and BPEL, it is difficult to implement the transformation of messages protocols and
dynamic routing.
Some of the original message middleware have been transferred into ESB products; such
message middleware and data bus are applied more in the original EAI application inte-
gration. SOA integration is based on Web Services and regards WS as its basic manage-
ment unit. Position of a service is about how to implement the business logic as a set of
mutually independent self-described and interoperable entity.
SOA is concerned about the whole life cycle of service through which the business
value would be achieved, while ESB is concerned about service intermediary and service
integration, which is the infrastructure of SOA. SOA has two core components, one is ESB
and the other is BPEL, and while ESB is infrastructure, BPEL is the service integration
driven by a business process. Without SOA, ESB will lose its connected services and is
only a bus with no value.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 245

The early establishment of SOA did not contain a large and complete ESB. However,
business problems should be given attention and solved using SOA. These business ser-
vices will create business value, and the services can be assembled to dynamically solve
changeable business requirements, making the user’s business flexible and diverse. In the
process of assembling, an ESB can be considered to connect these services together.
ESB requires a certain form of service-routing directory to route service requests.
However, SOA may also have a separate business service catalog most basic form of which
may be a design service directory used to achieve the reuse of services throughout the
development of the organization. Web Services vision places an Universal Description
Discovery and Integration (UDDI) directory in both business service directory and
­service-routing directory, thus making it possible to dynamically discover and call ser-
vices. Such a directory can be considered as part of the ESB. However, the business service
directory may be separated from ESB until such a solution has become widespread.

10.2.4  Industry Benefits from Implementing SOA

SOAs are built using a combination of industry standards and industry best practices.
The most important aspect of an SOA is the notion of modular components. Components
are interconnected through a common shared bus, often referred to as an ESB. The
ESB acts as the communication platform that allows the components to communicate.
Components can be written in a plug and play fashion to subscribe to the data of inter-
est on the bus, and publish data and provide services for use by other components.
The second key attribute of an SOA is the use of well-defined data formats with well-
defined semantics so that the components on the bus can understand the data and use it
appropriately.
In developing an SOA, each service should have a well-defined and published inter-
face. There should be very loose coupling between clients and services and the service
should be independent of the implementation technology. There should also be service
levels (QoS) defined for each service. Often the services are exposed to the ESB via Web
Services standards [7].
There are a number of benefits from implementing an SOA. Not only will SOA improve
the code reuse that can enable lots of savings, it will also give information about business
processes and how the business can be improved.
There are six benefits of SOA given in the following tables from the commercial
perspective:

1. The innovation will increase the speed of loose coupling of applications.

2. Cost of software development will be reduced as loose coupling allows smaller
projects to be implemented.
3. Alerts can be triggered to deal with real-time business events where help or
change is needed in the organizations.
4. Unstructured processes (i.e., activities outside the scope of automation) can be
integrated with the structured processes, which are called convergence.
5. Collaboration tools will be integrated with portals to increase productivity by
exposing services of role-based user workflow.
6. It is easier for loose coupling to update applications to new rules in order to evolve
the whole processes.
246 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

10.3  Research to Improve the Stability of SWIM Based on SOA

As key implementation of information exchange and data sharing in an aviation network,
SWIM provides an integration of various resources from the network, including airspace
management, flow management, air traffic management, surveillance management, and
aircraft system to manage correspondence, GPS, surveillance data, weather, and geo-
graphical information [3]. Thus there are several varieties of application servers deployed
in SWIM (Figure 10.4).
The SOA infrastructure is required to be built with great availability, allowing for opera-
tion of civil aviation, to weaken the influence of unpredictable conditions such as emergent
network delay, long response time, and reduction of QoS. All countries set up SWIM-SUIT
[8] based on a reliable network infrastructure.

10.3.1  General Introduction to Releasing the Thread Blocked of Application Server

The resilient SOA system is required to build SWIM for civil aviation. Blocked threads on
an application server is a major problem in building a resilient SOA, as shown by a large
number of practices [9].
As the main platform to deploy civil aviation business services, application servers are
an important infrastructure of SWIM used to implement service logic. Obviously, the sta-
bility of application servers is extremely important for a civil aviation air traffic control
system. However, numerous servers are distributed irregularly and often require thread
remote calls.
The solution of the problem requires using the present framework of SOA and avoiding
vast modification of the whole system. The manageability of the adjustment and ease of
implementation must be ensured even involving the configuration adjustment and a small
amount of refactoring SOA.

Airspace Flow Traffic Separation Aircraft

management management management management and others

SWIM

Common data transmission (CDT)

A/G Comm Navigation Surveillance Weather

World
(ACARS, (GPS, (ADS-B, (WARP,
(TAWS,
VDL2, VOR, radars, TCWF,
EFB,
ATN, TACAN, TCAS, OASIS,
DTED, etc.)
FANS, etc.) LORAN, etc.) ASDE, etc.) TDWR, etc.)

FIGURE 10.4
Architecture of SWIM based on SOA.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 247

This section presents two methods—adjusting the physical system deployment and
application servers:

1.
Adopting aggressive timer settings: It is the key step to speed up the average service
of every thread by dynamically changing timeout probability and timeout value.
Queue model reveals the direct influence of timeout and average response time
on system and has Remote Method Invocation (RMI) processed in 30 s before EC3-
abend happens by further calculation [10]. Thus, aggressive timer settings based
on the queue model will release blocked threads on some levels to strengthen
stability and productivity of SOA.
2.
Optimizing the deployment of SOA services or applications: In this method business
applications are deployed with tight coupling in the chain of synchronously inter-
connected components in the same physical system. Deployment of business
applications that are distributed in the network and share vast synchronous com-
munication aim to reduce RMI and improve SOA.

10.3.2 Adopting Aggressive Timer Settings: Methods Based on

Self-Contained Queue Model
10.3.2.1  Application of Aggressive Timer Settings
It generally takes much longer to use a passive timer to time than the average response
time. Dispatch timeout is set to 300 s while the process only takes 1–2 s. As Figure 10.5
shows, if a normal average response time of a service request is 3 s, it is still a small number
of requests causing RMI timeout that interferes with the speed rate of the server seriously,
even though RMI timeout value is averagely 30 s.
The coming threads are only waiting in line for deployment by server’s managed tasks
as the speed of the application servers is not as fast as that of coming threads. Provided
that all managed tasks are blocked, servers cannot work. In this case, all programs in line
will be out of time and fail. This will grandly affect stability and practicality of SOA.
In the WLM queue model, throughput of the servers will be increased as the simply
quantifiable aggressive timer sets the timeout value to 10 s. However, this timer could

Device 1
Average response
time: 3 s Device 2
(95%)
WT …

WT Device n
WLM queue
WT
...... Device 1

WT pool RMI timeout: 30 s Device 2

(5%)
...

Device n

FIGURE 10.5
Influential factors of service rate.
248 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

cause the end of part of service requests and add failures. Thus, service efficiency of serv-
ers will be weakened. As a result, the timeout value of an aggressive timer should be set
based on data dynamic of timeout probability in WLM queue model to speed up the ser-
vice of every thread and reduce the frequency of abnormal end of EC3.
In setting the aggressive timer, its attributive timeout probability belongs to continu-
ous data. Thus, it makes no sense to save and process data directly. In practical operation,
it is an effective way that the timeout probability value can be discretized into several
grades.

10.3.2.2 Implementation of Discretization of Timeout Probability

Attribute of Queue Model by KDD Language Field
With respect to the convenience of calculation in algorithm design of queue scheduling,
timeout probability should be processed in discretization in the first place. At present,
researchers came up with various discretization algorithms for continuous data such as
finite element method, finite difference method, the discretization algorithm based on
clustering, etc.
The procedure of discretization method is listed as follows [11]:

1. Attributive classification: Confirm the quantity of required discrete values, then give
the standard values of discrete radius, lower threshold, upper threshold, radius
of error (ε-neighborhood), and membership degree of discrete values. Generally,
these numbers can be offered by users or experts. Taking an example of timeout
probability, use the language of the five discrete values: very low, low, general, high,
and very high, and the corresponding standard samples of a1 = 10, a2 = 30, a3 = 50,
a4 = 70, and a5 = 90 (unit “%”), respectively. The radii of error are set to r1 = 2, r2 = 2,
r3 = 2, r4 = 2, and r5 = 2. The membership of discrete values can be given by the user
(or expert).
2. A continuous numerical u mapping to discrete values has two problems. First, if
u does not fall in the crossing range, it can be directly mapped to the correspond-
ing discrete values. Second, if u falls within the range of the intersection of ai and
ai+1, then U (nonstandard vector of u) can be calculated using a different formula
(wherein ai for the ith interval of standard sample points, li is the length of the
area, Ai is the standard vector of the area, and Aa is the standard vector of the adja-
cent area [may be either Ai+1 or Ai−1]). After that, it can calculate location according
to the minimum measure values of U and Ai, Ai+1, and Ai−1, which are calculated
using Hamming distance.
The theory is precise and simple for practical calculation, not only avoiding the
confusion of boundary value under the condition of a small amount of data, but
also taking account of dynamic data, which is suitable for this study.
3. Algorithm of discretization of continuous attribute based on language field theory: The
discretization method refers to defining lingual variables and their values, and
which determine to the interval boundaries. With a great amount of data existing
in the operating system; the database cannot load the extra data. Generally, upper
and lower threshold values can be confirmed by experts as the range of attribute’s
value. If the count of discrete value that is defined by user is five, then a1, a2, a3, a4,
and a5 are the standard samples. As shown in Figure 10.6, using a3 as an example,
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 249

r1 r2 r3 r4 r5

a1 a2 a3 a4 a5
L[1] L[3] L[5]
L[2] L[4]

AR[1,1] AR[1,2] AR[2,2] AR[3,2] AR[4,2]

AR[2,1] AR[3,1] AR[4,1]

Lower Upper
threshold threshold

FIGURE 10.6
The range of standard sample for continuous attribute discretization algorithm.

r3 will be the error radius of a3. If the attribute’s value is in the scope of [lower
threshold, AR[1,1]], [AR[i,2], AR[i + 1,1]], [AR[4,2], and upper threshold](i = 1,2,3). It
is a standard sample, beyond that will be a special sample [11].
The algorithm is illustrated as follows:
a. List is an ascending order of the attribute value table without repetition,
solving AR[i,1], AR[i,2], O[i], and L[i] and setting up their membership
b. For (i = 1; i < 5; i = i + 1);
c. List. Find Nearest (O[i]);
d. t = the value at the location pointed to by the pointer in the List;
e. If t ≤ AR[i,1], then Li[i,2] = AR[i,1], List. Next; else go to g);
f. If List.EOF = true, then Li[i + 1,1] = AR[i,2] and go to i);
g. If t ≥ AR[i,2], then Li[i + 1,1] = AR[i,2],List. Prior; else go to i);
h. If List.BOF = true, then Li[i,2] = AR[i,1] and go to i, else solving U by interpolation
formula, and determining Li[i + 1,1] and Li[i,2] by the value of Hamming
distance;
i. Continue for;
j. The end.
Using this algorithm can get boundary value as Li[i,2] and Li[i + 1,1], and i = 1,2,3,4.
WLM queue model automatically adjusts the levels of timeout values of the aggres-
sive timer referring to the previous dynamically changing data of timeout prob-
ability to weaken the influence of timeout on servers’ throughput. When timeout
probability increases to the error radius of critical value, timeout values can be
upgraded and added to improve service. Very high timeout probability leads to
servers getting blocked significantly. Servers will get rid of threads in line and
release resources to get over the block as soon as possible.
4. Analysis of the simulation experiment: If the arrival time of WT in queue is random
vector, it often follows the Poisson distribution. Each WT service time obeys expo-
250 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

nential distribution parameter to 1/EX. Among them, EX is the average service

time. If set at the average service time of 3 s, the distribution parameter is a third.
There are two kinds of simulation experiments: The first situation is the varia-
tion of timeout probability when the equipment is in normal condition. The second
situation is the variation of timeout probability when the equipment is out of order
on some scale. The result mainly analyzes the variation of server’s throughput
on condition of simple aggressive timer, passive timer, and probability dynamic
changes in levels of aggressive timer.
Experiment setting: If there are 40 WT pools and 20 referred-access equipment, the
arrival time of WT that WLM inputs obeys the Poisson distribution, and every ser-
vice time obeys exponential distribution (set EX = 3.0, distribution parameter = 1/3).
The simulation experiment adopts C# (VS.NET 2010) design implementation.
Random sequence using GUID as seeds makes its randomness more accurate.
Experiment 1: Variation of the throughput and timeout of WT probability on the
condition that all equipment work normally. Timeout values are divided into
four conditions to study according to the same group of random WT sequence:
(1) timeout = 30 s, (2) timeout = 60 s, (3) timeout = 120 s, and (4) timeout = 300 s. As
Figure 10.7 shows, on the condition that equipment and servers work normally,
the throughput will increase with timeout values decreasing, and meanwhile the
timeout probability will also increase a little. Thus, the experiment proves that
setting timeout does not exert an obvious influence on server’s throughput on the
condition of the equipment’s normal working.
Experiment 2: Part of the equipment breakdown. The breakdowns that cause time-
out of WT in servers are divided into two types: one is the breakdown of the
accessed equipment, the other one is the breakdown of correspondence threads
(router breakdown). The experiment assumes that the breakdown occurs to 10%
of equipment suddenly at the 300th second after the server starts its service and
results in the timeout of all WT accessing the equipment. The number of WT queu-
ing up in working thread pool increases due to the timeout of part of the equip-
ment, causing that the space to use in the pool shrinks, which results in chained
responses breaking down the normal procedure. Meanwhile, numerous time-
outs of WT reduce the server’s throughput and affects the stability of the whole
SOA framework. This experiment studies four conditions according to the same
group of sequence of WT from WLM: (1) passive timer setting timeout = 300 s; (2)
aggressive timer setting timeout = 30 s; (3) aggressive timer setting timeout = 60 s;
and (4) aggressive timer setting timeout = 120 s. The result is shown in Figure 10.8
that the average throughput decreased as the passive timer sets timeout = 300 s
and the average service rate is below 1 WT/S after 20  min, which slows down
the response speed of the whole server and reduces the resilience and stability of
the SOA system. In the other case, the throughput remains stable with a simple
aggressive timer, but timeout probability is too high to retain normal WT, which
means the efficient throughput is reduced sharply.
Thus, it can be seen that WLM requires a method to keep the throughput stable
and keep the timeout probability from growing too high. The experiments prove
that neither simple aggressive timer nor passive timer can realize the requirement.
So this chapter uses the language field theory in data mining method to set the
timeout value level of application server dynamically. Its principle is based on the
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 251

10

9 Timeout = 30 s
Timeout = 60 s
8 Timeout = 120 s
Timeout = 300 s
7
Average throughput (WT/s)

6.1
6 5.93
5.86
5.81
5

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time (s)

100
Timeout = 30 s
90 Timeout = 60 s
Timeout = 120 s
80 Timeout = 300 s

70

60
Error WT (%)

50

40

30

20

10 8.01%
4.01%
1.65%
0.41%
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time (s)

FIGURE 10.7
(See color insert.) Four kinds of throughput and timeout probability curve on the condition of trouble-free
equipment.
252 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

10

9
Timeout = 30 s
8 Timeout = 60 s
Timeout = 120 s
Average throughput (WT/s)

7
Timeout = 300 s

4
3.62
3
2.64
2
1.44
1
0.74

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

100
Timeout = 30 s
90 Timeout = 60 s
Timeout = 120 s
80 Timeout = 300 s

70

60
Error WT (%)

50

40
34.56%
30 29.95%
29.33%

20 20.03%

10

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

FIGURE 10.8
(See color insert.) Variation curves of 10% of equipment breakdown at the 300th second during the experiment.

timeout probability of continuous attributes that is divided into five grades: very
low, low, general, high, and very high. The corresponding timeout probability is set
to 5%, 25%, 50%, 75%, and 95%, and the language value of the membership degree
is given. For example, the membership of general is A3 = [1.0, 0.8, 0.5, 0.2, 0] and its
threshold limit is 0% to 100%, as is shown in Figure 10.9, when set r = 5%, because
the timeout probability value is a continuous change, in the crossing range will
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 253

ri ri+1

ai Crossing ai+1
range

AR[i,1] AR[i,2]
AR[i + 1,1]

Lower Upper
threshold threshold

FIGURE 10.9
Crossing range.

not be able to accurately correspond to a discrete value. For example, when the
probability falls within the range of cross between 10% and 20%, which can be
divided into levels, we will need to use the language field theory to calculate. With
the increase of the amount of data, the boundary value will be more and more
accurate. This can be the continuous value of timeout probability corresponding
to the discrete level of the timer. Use of a dynamic timeout value will satisfy the
stability of the server throughput and try to avoid the rapid increase of timeout
probability.

Experiment 3: Application server and access equipment work normally in initial
state, then 10% of the equipment breaks down causing timeout of part of WT.
This simulation experiment contains three conditions according to the same
random sequence of WT: passive timer, simple aggressive timer, and self-con-
tained dynamic timer. It is obvious that the throughputs of the simple aggres-
sive timer and self-contained dynamic timer are much higher than that of the
passive timer. Next, sample 10 groups of sequences of WT randomly and every
group of the same sequence is simulated with the simple aggressive and self-
contained dynamic timers to work out the average throughput of WT (as shown
in Table 10.2).

Consequence 1: The sample means of “WT throughput” are 2.489 and 3.663. The
sample variances are Ss2 = 0.020 and Sa2 = 0.021. The average throughput of WT of
the self-contained dynamic timer is higher than that of the simple aggressive
timer, and the stability of both near the sample mean is essentially the same.

TABLE 10.2
Contrast of WT Throughput (WT/S) between Adaptive Dynamic Timer and Simple
Aggressive Timers
Group of Experiments 1 2 3 4 5 6 7 8 9 10
Self-contained dynamic timer 3.62 3.53 3.69 3.83 3.87 3.49 3.44 3.67 3.71 3.78
Simple aggressive timer 2.46 2.42 2.69 2.29 2.73 2.60 2.36 2.40 2.46 2.48
254 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Consequence 2: Comparison between paired data on the same condition can use
the confidence of alpha = 5% of the t test for hypothesis testing. It is proven that the
throughput of self-contained dynamic timer is sharply higher than that of simple
aggressive timer.
5.
The experimental conclusion: This section presents the blocked threads of appli-
cation servers in SOA and provides a method that puts forward the strategy of
the dynamic aggressive timers. Timeout probability with continuous attributes
and language field theory can transform discretization processing into the lev-
els of timers. This method solves the problem of calculating the critical value of
implementation. This method avoids the problem of level boundaries of chaos
under the small amount of data, and does not increase the load with the large
amount of data.

Although the self-contained queue model releases the blocked threads of application
servers in SOA on some level, an excess of improper thread remote procedure calls still
causes the block. In the next section we will focus on dynamic adjustment of application
programs of SWIM in civil aviation and deployment of servers, as well as put forward
methods of reducing thread remote procedure call. These methods will cooperate with the
self-contained queue model to improve the resilience of SOA infrastructure.

10.3.3  Optimizing the Deployment of SOA Services or Applications

In fact, it is difficult to achieve the thread centralized application server program in a
truly comprehensive way according to the features of the hardware infrastructure of civil
aviation SWIM. This chapter discusses a class of short-term solutions, which can be eas-
ily applied to existing SOA framework without changing (or with just a few changes) the
overall architecture. Such solutions involve only configuration adjustments and other
minor optimization, and hardly require reconstruction of the SOA to carry out [9].

10.3.3.1  Optimizing and Deploying Tightly Coupled SOA Services

The basic idea of this method is to optimize the deployment of the tightly coupled network
services in the SWIM or applications on physical systems. Tightly coupled SOA services
refer to the relationship between multiple services called synchronously; when a block-
age happens to the service the client calls, the service will not continue until it receives
a response. In general, many tightly coupled SOA services often appear in a transaction,
thereby forming a synchronous interconnect assembly chain in SOA.
The short-term solution is to deploy tightly coupled business applications (or services) in
the synchronous interconnect assembly chain together as possible. The focus is on busi-
ness applications deployed on a separate server, but sharing a lot of synchronous com-
munication with each other (i.e., tightly coupled to each other). By redeploying to the same
server, the overall condition and stability of the SOA are improved.

10.3.3.2  Advantages of Optimized Deployment

The two main advantages of tightly coupled business applications (or services) in SOA
deployed together are as follows.
First, the in-process communication protocol can be taken advantage of to strengthen
direct communication between applications when they are deployed in the same host.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 255

The in-process communication can avoid all the cost remote service invocations (includ-
ing serialization, encryption, traversing the network stack, and network latency), so this
method can bring significant performance improvement for the entire system.
Second, it can reduce pressure on the resource requirements to deploy the tightly cou-
pled business applications together, especially the thread-level resources. Synchronous
service call through the local communication protocol in local in-process typically gen-
erates less pressure on local servers than deployment from the remote communication
protocol does.

10.3.3.3  Optimized Deployment under High Workloads

Such synchronization-dependent applications deployed on a different server typically
generate HTTP or RMI communications connections (Figure 10.10). For example, an exter-
nal client calls application A in the server 1. Their work request will be sent to the hosted
task of the server 1. However, because the communications between the applications A
and B are synchronized (e.g., RMI/IIOP), the hosted tasks distributing work to application
A in the server 1 must be blocked until it receives a response from the application B and
when blocked, hosting tasks are all waiting in line.
If the server suffers from a heavy load at this time, only a small part of hosted task
resources are available, so more and more hosted tasks are blocked on synchronous remote
service calls, and the deploying ability of server 1 would be further reduced, resulting in
the decline of server 1 in the rate of overall service quickly. When the service rate is lower
than the arrival rate of new tasks, the server will not keep up, so the new task will be wait-
ing in line. The worst outcome is that all hosted tasks are blocked, and the server cannot
perform any work. In the case that this problem persists for a long time, queue tasks and
distributed tasks will run overtime and eventually fail.
The solution to this problem is to deploy tightly coupled business applications on the
same server, which apparently reduces pressure from applications on resources (Figure
10.11). Application A can use the local in-process protocol to call service B as local calls
directly, which avoids occurrence of a remote call and a series of problems related to the
remote delay. This will reduce the blocked managed tasks, so that the server can maintain
a high rate of service, effectively avoiding the queue growth and preventing timeout of the

Server 1 Server 2

Application A Application B

Communication channel

FIGURE 10.10
HTTP or RMI communications connection of a server to another server.
256 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

Server 1 Server 2

Application A
Application B

Application B

Communication channel

FIGURE 10.11
Deploying tightly coupled business applications on the same server.

task deployment, which in the end further reduces instability in the environment of the
SOA to give a more stable and more resilient SOA system.

10.3.3.4  Preferable Service Deployment Solution Based on Clustering Analysis of Log

Due to the characteristics of civil aviation applications, the SWIM application deployment
is in certain particularity, and services and applications of SWIM are also in a constant
process of change. Therefore, this chapter presents a log-based cluster analysis method to
find a preferable system deployment program.
The first step is data preparation—to extract attributes of service (or application) objects
in log, which involves the following: (1) conform the sets of all services (or applications),
(2) obtain each service server address, (3) identify the total number of times each applica-
tion calls other applications, (4) record RMI for each service that occurs (including the
times and RMI objects); (5) set a moveable for each application that cannot be migrated in
order to ensure the feasibility of the optimization program, since there is a certain limita-
tion for the migration position of applications in the civil aviation network.
The second step is service (or application) set division. A specific division method is
used if the number of application programs is n, the number of servers is k to divide clus-
ters, and the initial state of each cluster is the application set Ak of every server (excluding
nonmigrated applications).
In the third step, the set will be clustered by a typical algorithm K-means, calculation
of the distance between each application from the times of calls and the times of RMI
occurrence.
Each cluster (≤N) must meet the following two conditions: (1) comprise at least one object
and (2) each object must belong to only one cluster. Then to improve the data similarity
between clusters, an object is moved from one division to another based on the differ-
ence between the clusters by an iterative relocation algorithm based on the initial division.
A general guideline of a preferable division is the objects in the same class are possibly
close or similar, but the objects in different classes are away or different.
Suppose there are n samples of data, divided into k clusters (k ≪ n) for t iterations, the
time complexity of the serial K-means algorithm is O (n * k * t). Obviously, in order to reach
the global optimal, division-based cluster will usually require exhausting all possible
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 257

divisions, but K-means algorithm in actual use ends generally based on local optimum. The
K-means dividing method that converges faster is more suitable for small and medium-
sized grouping of data elements.
Usually parallel computing is recommended for solving data problems with such a large
amount of computation, and K-means is a parallelized algorithm. Therefore, this chapter
proposes the Hadoop cloud computing platform to solve complex computing problems in
the civil aviation SWIM system. For example, in the Hadoop environment, data stored in
the Hadoop distributed file system (HDFS) file is automatically divided into data blocks of
64 K. In the map phase, a data block is dealt with using a map task and the entire data file
is assigned to a different node. As each node can complete i tasks and j nodes participate
in the parallel computation, time complexity of parallel K-means algorithm is O (n * k * t */
(i * j)). Thus, in theory, the time efficiency of the post-parallel K-means algorithm is indeed
greatly improved.

10.4  Cluster Analysis on the Hadoop Cloud Computing Platform

10.4.1  Introduction of Hadoop Cloud Computing Platform [12]
Hadoop is a cloud computing platform that can be easily developed and processes lots of
data in parallel. Its main features include strong expansion capacity, low cost, high effi-
ciency, and good reliability. First, HDFS of Hadoop uses the M/S architecture. An HDFS
cluster is composed of a management node (name node) and a certain number of data
nodes (data nodes). Each node can be an ordinary PC. In use, the stand-alone file system
and HDFS are very similar, and HDFS can also build a directory; create, copy, and delete
files; and view file contents. However, its underlying implementation is to split the file into
blocks, and these blocks are stored on different nodes. Each block can also be copied and
stored on different nodes in order to achieve fault tolerance. The management node is a
central server that is responsible for managing name space (namespace) of the file system
and access to client end on files. The data nodes in the cluster manage the data store on the
node. The management node is the core of HDFS, which records how many blocks are cut
by maintaining a group of data structure. The blocks can obtain from those nodes impor-
tant information like the condition of each data node. Besides, MapReduce of Hadoop is an
efficient distributed programming model, and a way of processing and generating a large
number of data sets.

10.4.2  Exploration of Hadoop Clusters and Internet

10.4.2.1  Basic Principles of Operation of Hadoop Clusters
There are three parts of main task deployment in Hadoop: client machine, the master-, and
slave nodes. The master node is responsible for the supervision of two key functional mod-
ules, such as HDFS and MapReduce. When the job tracker uses MapReduce to do parallel
processing, the name node is responsible for monitoring and dispatching HDFS. The slave
node is responsible for most of the machine’s running and working on all data storage and
command computing. Each master node is a data node but also a daemon to communicate
with its master node. Daemon is affiliated with the job tracker, and data nodes are affili-
ated with the name nodes [13].
258 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

The client machine is assembled with all the cluster settings on Hadoop, but without mas-
ter nodes and slave nodes. Instead, it is the client machine that loads data into clusters and
submits description of data processing to MapReduce, and gets it back or checks out the
results after work. In a small cluster (approximately 40 nodes), it often happens that a single
physical device handles multiple tasks such as job tracker and name node simultaneously. As
the middleware of a large cluster, a separate server is generally used to handle a single task.
There is no virtual server or management layer in the real product cluster to avoid
unnecessary performance loss. Hadoop works best in Linux, operating directly with the
bottom hardware facilities, which shows that Hadoop actually works directly on the vir-
tual machine. This is superior in terms of cost, ease of learning, and speed.

10.4.2.2  Structure of the Typical Hadoop Cluster

The structure of a typical Hadoop cluster is shown in Figure 10.12. A series of racks are con-
nected by a large number of rack switches with rack servers (not the blade servers), which is
usually held out with 1 or 2 GB broadband. Ten gigabytes bandwidth is not common but can
significantly increase the density of the CPU core and disk drives. The last rack switch layer
connects a number of racks simultaneously together by the same bandwidth and forms
clusters. Lots of servers would become the slave nodes with disk storage machines, CPU,
and memory (DRAM). Also some machines would become the master nodes and these
machines with a small number of disk storage machines own a faster CPU and more DRAM.

10.4.2.3  Hadoop Workflows

As business users have large amounts of data to be analyzed and processed quickly,
Hadoop separates massive data and dispatches the data to the computer to do the parallel

Hadoop server roles

Clients

Distributed data processing Distributed data storage

Map reduce HDFS

Secondary
Job tracker Name node Masters
name node

Data node and Data node and Data node and

task tracker task tracker task tracker
Slaves
Data node and Data node and Data node and
task tracker task tracker task tracker

FIGURE 10.12
(See color insert.) Hadoop server roles.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 259

process. For example, if the user has a large number of data files (such as e-mails to the
customer service department): The number of times refund appears has to be counted in
time, which is a simple task of word count. Customers load data into clusters (file.txt) and
submit a description of the data analysis, - word count. Then clusters will store the result in
a new file (results.txt), and the client reads the result document. Typical workflow is shown
as follows:

• Load data into the cluster (HDFS writes)

• Analyze the data (MapReduce)
• Store results in the cluster (HDFS writes)
• Read the results from the cluster (HDFS reads)

The Hadoop cluster does not work until the data is input into it, so a large file.txt has to be
loaded to the cluster above all, and the aim is fast parallel processing of data. To achieve
this goal, as many machines as possible have to work simultaneously. Finally, the client
will divide the data into smaller modules that are dispatched to different machines of the
cluster. The smaller modules the data are divided into, the more machines can do parallel
processing of data. Since these machines may be out of order, a single block of data has
to be processed simultaneously on different machines in order to avoid loss of data. Each
block of data would be loaded over and over again on the cluster. Generally, the default
settings of Hadoop are that each block of data is repeatedly loaded three times and this
feature can be set via dfs-replication parameters in hdfs-site.xml file.

10.4.3  MapReduce Data Flow [13]

The MapReduce client includes three working elements: input data, MapReduce program,
and configuration information. Hadoop divides the job into map tasks and reduce tasks.
The job execution process employs two types of nodes: the job tracker and numerous
task trackers. The function of job tracking is to coordinate all the tasks through the sched-
uling task running on the system. The function of task tracking is to send the progress
report of the tracking task, which records the progress of each job. If the task fails, the job
tracker can rearrange the task tracking.
Hadoop divides tasks into fixed size pieces in the MapReduce. These pieces are known
as input splits or simply splits. Hadoop creates a map task for each split to execute each
record in a user-defined mapping function.
If there are a lot of split blocks, in fact, the time of processing each split is less than that
of processing the whole task. A faster machine will be able to process proportionally more
splits over the course of the job than a slower machine. If we process the splits in parallel,
the processing is better load-balanced if the splits are small. With a split into small par-
ticles, even if the machines are identical, failed processes or other work carried out simul-
taneously make load balancing desirable and the quality of the load balancing increases.
On the other hand, if splits are too small, then the total job execution time will be domi-
nated by the overhead of managing the splits and of map task creation. For most jobs, a
good split size tends to be the size of a HDFS block, which is 64 MB by default, although
this can be changed or specified for the cluster (for all newly generated files) when each
file is created.
Hadoop tries its best to run the map task on a node where the input data resides in
HDFS, which is called the data locality optimization. We now see why the optimal split
260 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

size is as small as the block size; that is it ensures the maximum size of input that can be
stored on a single node. If the split was divided between two blocks, unfortunately any
HDFS node can store both blocks, so some of the split would have to be transferred to the
node running the map task through the network, and it is obvious that this is less efficient
than running the whole map task by using local data.
A map task will output information directly to the local disk, rather than to HDFS. The
map output process is an intermediate output. Its specific process is reducing the process-
ing tasks to produce the final output, and once the work is complete the map output can
be thrown away. So it cannot be stored in HDFS to avoid wastage. If a node fails before the
map has been already output, Hadoop will automatically re-run the task to create a map of
the output mapped on another node.
Reduce tasks do not have the advantage of data locality. That is the input to a single
reduce task is normally the output from all mappers. For example, now we have a single
reduce task fed by all of the map tasks. Thus, the sorted map outputs must be sent through
the network to the node where the reduce task is running, they are the first to be combined
and then transferred to the user-defined reduce function. Typically, the output of reduce is
stored in HDFS to ensure the reliability. Therefore, the reduce output consumes network
bandwidth, but only as much as a normal HDFS write pipeline.
The data flow for the general case of multiple reduce tasks is illustrated in Figure 10.13.
This diagram makes it clear why the data flow between map and reduce tasks is colloqui-
ally known as the shuffle, as each reduce task is fed by many map tasks.

HDFS Split 0 Split 1 Split 2 Split 3

Read Read Read Read

(input format) (input format) (input format) (input format)

Mapper Mapper Mapper Mapper

Map Map
task a 1 b 1 c 1 c 1 a 1 c 1 b 1 b 1

Partitioner Partitioner Partitioner Partitioner

Shuffle and sort Shuffle and sort Shuffle and sort

Reduce a 1 a 1 b 1 b 1 b 1 c 1 c 1 c 1 Reduce
task
Reducer Reducer Reducer
Write Write Write
(output format) (output format) (output format)

HDFS Part-0 Part-1 Part-2

FIGURE 10.13
(See color insert.) MapReduce data flow.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 261

10.4.4 Design and Implementation of Parallel K-Means Clustering Algorithm

Based on Hadoop Platform
In the design of Hadoop parallel algorithm, the user’s main job is to design and implement
Map and Reduce functions, including input and output <key, value> key type and Map
and Reduce logic, and so on [14].
As can be seen from the serial K-means algorithm, the main computational work is to
assign each sample to its nearest cluster, and operations to assign different samples are
mutually independent for which parallelization is required. In iteration, the algorithm
performs the same operation, and the parallel K-means algorithm (Parallel K-Means),
respectively, performs the same Map and Reduce operations in iteration.
Put K samples selected randomly as a focal point and save it into a file of HDFS as a
global variable. Then iteration consists of three parts: Map function, Combine function,
and Reduce function.
The following are the detailed descriptions of Map function, Combine function, and
Reduce function.

10.4.4.1  Function Map()

The <key, value> pair input by Map function is the default format of the MapReduce
framework, that is, key is offset that the current sample relates to the file starting of the
input data and value is a character string composed of coordinate values of each dimen-
sion of the current sample. First, parse out the value from each dimension of the current
sample from value; then calculate the distance between that and k center points and find
the nearest cluster subscript; and finally, output <key’, value’>, where the key’ is the
nearest cluster subscript and value’ is a character string composed of coordinates of each
dimension of the current sample. Function pseudocode is

Map (<key, value>, <key ’, value’>)

{
Analyze sample object from the value, denote instance;
Auxiliary variable MinDis is initialized to the possible maximum value;
Index is initialized to −1;
For i = 0 to k − 1 do {
dis = distance between the center of the ith and instance;
if dis <minDis {
minDis = dis;
index = i;
}
}
Refer index as a key ’;
Refer coordinate value of each dimension as value ’;
Output <key ’, value’>;
}

In order to reduce the communication cost and the amount of data transferred during the
iterative algorithm, PK-means algorithm design Combine after the Map and merge the
output data in local after Map function processing. Since the data output after each Map
operation is always stored on the local node, each Combine operation is performed locally,
and the communication cost is small.
262 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

10.4.4.2  Function Combine()

In the Combine function input <key, V> pair, key is the cluster (also called grouping)
subscript, V is a character string list consisting of dimensional coordinate values of every
sample assigned to the cluster whose subscript is key. First, analyze the dimensional coor-
dinate values of each sample from the list, and corresponding coordinate values of each
dimension are added together, and record the total number of samples in the list while.
Key’ is the cluster subscript in Output <key ’, value’> pair; value’ is a character string
that includes two parts of information: the total number of samples and the accumulated
coordinate values for each dimension and composition of the character strings. Function
pseudocode is

Combine (<key, V>, <key ’, value’>)

{

Initialize an array to store the accumulated value of the coordinates of each dimension,
and the initial value of 0 of each component;
Initialize the variable num, record the number of samples assigned to the same cluster,
and the initial value is 0;

While (V.hasNext ()) {

Analyze the coordinate values of each dimension of a sample from V.next () in;
The coordinate values of each dimension are added to the corresponding component in
the array;

num ++;
}
Refer the key as a key ’;

Construct a character string that contains num and information of all components in an
array, and refer this character string as value ’;

Output <key ’, value’>;

}

10.4.4.3  Function Reduce()

In the <key, V> input of the Reduce function, key is the cluster subscript and V is the inter-
mediate result from the transmission of each Combine function. In the Reduce function,
parse out the number of samples processed from each Combine and coordinate accumu-
lated values of each dimension of the corresponding node at first; then the corresponding
accumulated values of each dimension add correspondingly in separate way, which is then
divided by the total number of samples, that is, the new center point coordinate. Function
pseudocode is

Reduce (<key, V>, <key ’, value’>)

{

Initialize an array to store the accumulated value of the coordinates of each dimension,
and the initial value of each component is 0;
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 263

Initialize variable NUM, record the total number of samples assigned to the same cluster,
and the initial value is 0;

While (V.hasNext ()) {

Parse out the coordinate values of each dimension of a sample and the number of samples
num from V. next ();
The coordinate values of each dimension are added to the corresponding components
in the array;

NUM + = num;
}

Each component of the array is divided by NUM to get a new center coordinate;

Refer the key as a key ’;

Construct a character string containing the information of coordinate values of each

dimension of the new center point, and refer the character string as value ’;

Output <key ’, value’>;

}

Obtain coordinates of new center point according to the output of Reduce and update the
coordinates to the file on the HDFS; then, perform the next iteration until the algorithm
converges.
K-means algorithm is one of the most commonly used algorithms in practice, which
has an absolute advantage in dealing with a large amount of data and can achieve better
results. Based on the results obtained through clustering, application, or service that often
call each other will be assembled in a cluster (a server), and, of course, there will be some
applications frequently called by the server remotely assembled in this cluster. Apparently,
it is more reasonable that these applications are deployed in the cluster (or the server),
which can give advice dynamically on adjusting and optimizing application deployment.
It has great significance in solving complex computational problems to introduce a cloud
computing platform into the SWIM system. This section, taking the open source Hadoop
platform as an example, uses this platform for clustering computation in order to optimize
deployment of SOA-based applications in SWIM in each server to further improve the
stability of SOA and production availability.

10.5 Conclusion
This chapter presents the basic structure of ATN and the features of data information
that includes ADS-B, an important part of a new generation of ATM systems; CPDLC,
enabling pilots and controllers to carry out two-way communication directly via a data
link so that they can send a variety of requests and control commands; and FIS, supporting
the pilots to get all sorts of information during flight. Then, it explains the importance of
the relation between SWIM and ATN. SWIM’s standardized TCP/IP protocol has reached
264 Aeronautical Telecommunications Network: Advances, Challenges, and Modeling

a higher bandwidth so that the information sharing and collaborative decision-making

can be realized through a standardized interface. FAA plans to achieve full integration of
ATN and SWIM in the long term. Also the structure and logic layers of SOA architecture
are analyzed, as it consists of operating system layer, enterprise components layer, services
layer, business process composition layer, access or presentation layer, integration (ESB),
and QoS. As the core infrastructure of SOA, ESB is referred as a service intermediary
providing functions of data transmission, message protocol conversion, and routing in the
integration of heterogeneous systems. This chapter presents ESB’s concept and working
principle.
This chapter also focuses on the transmission control of ATN based on SOA. As the resil-
ience of SOA has a key effect on ATN and SWIM, this chapter presents two approaches to
improve network resilience based on SOA architecture:

1.
Optimize and improve the deployment of SOA services or applications: The center of
such a solution is a close coupling of business applications deployed on the same
physical system as much as possible in synchronous interconnect assembly chain.
It is responsible for deploying business applications dispersed in the network but
shares a large amount of synchronous communication, which reduces the RMI
and provides a great benefit for business application programs so as to improve
the overall condition and stability of SOA.
2.
Using active timer program: The key to solving congestion problems is dynamically
changing the timeout and the timeout value to increase the probability of the
average service rate for each thread and to improve the service rate for the entire
server. The introduction of the queue model can intuitively show the impact of
timeouts and the average response time on the system; on the other hand, further
calculation can enable the system to deal with RMI within 30 seconds before EC3-
abend [5] occurs. Accordingly, the timer queue based on active server model can
solve the problem of congestion to some extent; the thread, thereby enhances the
availability of SOA and production stability.

This chapter also introduces the Hadoop cloud computing platform while dealing with
server deployment, including Hadoop common, Hadoop distributed file system, Hadoop
YARN, and Hadoop MapReduce, whose main features include strong expansion capacity,
low cost, high efficiency, and good reliability. And, it analyses the log of remote proce-
dure call and the working principle of Hadoop cloud computing platform, and introduces
MapReduce.
The clustering algorithm is adopted to analyze the system log and to dynamically
adjust process deployment. As can be seen from the serial K-means algorithm, the main
computational work is to assign each sample to its nearest cluster, and operations to
assign different samples are mutually independent, for which parallelization is needed.
In iteration, the algorithm performs the same operation, and the parallel K-means algo-
rithm (Parallel K-Means) respectively performs the same Map and Reduce operations in
iteration.
K-means algorithm is one of the most commonly used algorithms in practice, which
has an absolute advantage in dealing with a large amount of data and can achieve bet-
ter results. Based on the results obtained through clustering, applications or services that
often call each other will be assembled in a cluster (a server), and, of course, there will
be some applications frequently called by the server remotely assembled in this cluster.
ATN Transmission Control Algorithm Based on Service-Oriented Architecture 265

Apparently, it is more reasonable that these applications are deployed in the cluster (or
the server), which can give advice dynamically on adjusting and optimizing application
deployment.
The clustering algorithm makes for increasing network flexibility and availability and
reducing RMI and congestion.

References
1. ICAO DOC. Manual of technical provisions for the aeronautical telecommunication network
(ATN). 97052AN/956-2001, 2001, pp. 34–35.
2. Luckenbaugh, G., Dehn, J., Rudolph, S., and Landriau, S. Service oriented architecture for the
next generation air transportation system. In 2007 Integrated Communications Navigation and
Surveillance Conference, 2007, pp. 1–9.
3. Harkness, D., Taylor, M.S., Jackson, G.S. et  al. An architecture for system-wide information
management. In 25th Digital Avionics Systems Conference, Portland, OR, 2006, pp. 1A6 1–1A6 13.
4. Sajjad, W. An optimum architecture for SOA. Daffodil International University Journal of Science
and Technology, 9(1), January 2014, 47–48.
5. Travis, B. Section 3: Developing service-oriented architectures. http://msdn.microsoft.com/
en-us/library/aa302164.aspx. Visited October 2014.
6. Flurry, G. Exploring the Enterprise Service Bus, Part 1: Discover how an ESB can help you
meet the requirements for your SOA solution. 2008. http://www.ibm.com/developerworks/
library/ar-esbpat1/. Visited December 2014.
7. Waheed, S. An optimum architecture for SOA. Daffodil International University Journal of Science
and Technology, 9(1), January 2014.
8. Houdebert, R. and Ayral, B. Making SWIM interoperable between US and Europe. In 2010
Integrated Communications Navigation and Surveillance Conference, Herndon, VA, 2010, pp. C4 1–
C4 8.
9. Snehal, A., Alderman, R.G. Build a resilient SOA infrastructure, Part 1: Why blocking applica-
tion server threads can lead to a brittle SOA[C/OL], 2007. http://www.ibm.com/developer-
works/webservices/library/ws-soa-resilient/. Accessed December 2014.
10. Rica, W., Cleberson, C., Nguyen, K. et al. Problem symptoms in websphere for z/OS and their
resolution. In IBM, 2006, pp. 20–26.
11. Zhou, Y. and B. Yang. The discretization method and its realization of continuous attribute
based on language field theory. Journal of Computer Science (Chinese), 30, 2003, 63–66.
12. Apache Software Foundation. Welcome to Apache™ Hadoop®! http://hadoop.apache.org/.
Visited December 2014.
13. White, T. Hadoop: The Definitive Guide (1st edition), O’Reilly Media, Inc., Sebastopol, CA, June
2009, pp. 75–80.
14. Venner, J. Pro Hadoop. Apress, December 2009, pp. 27–30.
 90–110 130–535 kHz
NOB
30 kHz LORAN C 300 kHz

130–535 kHz 1,800–2,000 HF COMM

NOB
300 kHz LORAN A 3 MHz

3023 S&R 5680 S&R

HF COMM (2,850–22,000 kHz)
3 MHz 30 MHz

74.8–75.2 108 117.97 137

Marker Beacon VHF COMM
30 MHz VOR ILS 300 MHz

1,545– 1,559–
328.6–335.4 960 1,215 1,260 1,400 1,555 1,610 1,626.5 2,300–2,400 2,700–3,100
ILS Guide DME RNSS
slope
300 MHz 3 GHz

3,100 5,350–5,470 8,750–8,850 9,000–9,500 13,250–13,400

Surveillance radar MLS ASDE
3 GHz Radio altimeter 5,030–5,091–5,150 Airborne Airborne Precision 24.25– 30 GHz
weather doppler approach 24.65
radar radar radar

Aeronautical communication systems

Aeronautical surveillance systems

Aeronautical radionavigation systems

FIGURE 1.4
Communications radio navigation and surveillance bands.
 Frequency
3 Hz 30 Hz 300 Hz 3 kHz 30 kHz 300 kHz 3 MHz 30 MHz 300 MHz 3 GHz 30 GHz

Band ELF SLF ULF VLF LF MF HF VHF UHF SHF

100,000 km 10,000 km 1,000 km 100 km 10 km 1 km 100 m 10 m 1m 10 cm 1 cm

Wavelength

117.975–137 MHz
AMS (R) Sat Comms
2.85–23.35 MHz VHF 1,544–1,555 MHz AMS Sat Comms
HF mobile 1,645.5–1,646.5 MHz 14–14.5 GHz

Mobile JTIDS/MIDS
Communications

SINCGARS HAVEQUIK

ILS glide beacon

VOR/ILS localizer beacon

MLS
DME
Radio navigation
NDB Rad Alt

LORAN C LORAN A
Marker beacon GNSS
GNSS

Airborne weather Airborne weather

radar radar ASDR
SSR/ACAS 15.4–15.7 GHz

Surveillance
SSR/ACAS

Primary radar
Precision approach/ Airborne weather
ASDR radar radar

FIGURE 1.5
Aeronautical radio spectrum.
 Remote facility
edge base station Remote facility
edge base station
Hub

Remote facility Head

edge base station administrator
Hub Remote facility
block edge base station

Core
network
possible
ATC disaster
ATC control
Routing recovery site
options
triplication

Airport
Remote facility
edge base station Remote facility
edge base station

Hub Hub Hub

Remote facility
edge base station

Remote facility
edge base station Remote facility
edge base station

FIGURE 1.6
Hypothetical Aeronautical Telecommunication Network. (From ICAO, Aeronautical Telecommunication Network
(ATN): Manual for the ATN Using IPS Standards and Protocols, Doc 9896. September 2008.)

Application Application
layer layer
CSI/PS CSI/PS
convergence convergence

Peer-to-peer connections

Transport layer Transport layer

(TCP, UDP) (TCP, UDP)
Inter-domain Inter-domain
router router
Internet layer Internet layer Internet layer Internet layer
(IPv6) (IPv6, BGP4+) (IPv6, BGP4+) (IPv6)

Link layer Link layer Link layer Link layer

Local or Local or
intradomain intradomain
subnetwork subnetwork
Interdomain
subnetwork

FIGURE 1.7
Aeronautical Telecommunication Network/Internet protocol suite protocol architecture. (From ICAO,
Aeronautical Telecommunication Network (ATN): Manual for the ATN Using IPS Standards and Protocols, Doc 9896,
September 2008.)
 Air–ground link Handover to
Hand-over satelite link

Polar or ocean area

ATC #2

ATC #1
in
oma
e rd
Int ter
rou
Network domain #2
Network domain #1

FIGURE 2.1
ATN mobile communication framework.

HA in Europe
HA in Asia

HA in the
United States

Communication path
MR-HA tunnel
HA-HA signaling
CR
Correspondent
network

(a)

Home network
HA

CR
CN Correspondent
network
MR-HA tunnel Communication path via HA
MR-CR tunnel Communication path via CR
(b)

FIGURE 2.10
Typical air traffic services communication scenario making use of different route optimization protocols.
(a) Global HA-to-HA and (b) Communication router (Adapted from Bauer, C., Secure and Efficient IP Mobility
Support for Aeronautical Communications. Karlsruhe, Germany: KIT Scientific Publishing, 2013.)
 Radar data Conflict detection Flight plan
processing and alarm

A A
B B

Integral surveillance Integral surveillance

and control(I) and control(II)

FIGURE 3.1
Composition of ATC.

36 CLAM 36 STCA
CSN6201 M CSN6201 M
0056 V 0090 031 0056 V 0090
CAUC A319 CAUC A319

FIGURE 3.4
Conflict warning.

FIGURE 3.7
Overview of ISC.
FIGURE 3.8
Track.

FIGURE 3.9
Maps.

FIGURE 3.10
Flight plan window.
FIGURE 3.11
Flight strip.

Communication
satellite
VHF

RGS RGS
RGS

Ground
network
Users
Network management and
data processing center

FIGURE 4.1
The composition of ACARS.

FIGURE 4.2
ACARS receiving software acarsd 1.65.
 Key Intra-
distribution organization
network
VHF
ATSP

Intra-
organization
network
DSP
Airline

Ground–
Air–ground ground Intra-
network network organization
network
Security envelope Third-party
service provider

FIGURE 4.3
DSP-based security architecture.

Key
distribution Intra-
organization
network
VHF
ATSP

Intra-
organization
network
DSP Airline

Ground–
Air–ground ground Intra-
network network organization
network
Third-party
Security envelope
service provider

FIGURE 4.4
End-to-end security architecture.

FIGURE 5.7
The GS node model.
FIGURE 5.9
The GS MAC process.

FIGURE 5.10
The AC handoff child process.
FIGURE 5.14
Handoff scenario.

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS
Dashed line Dash dot line Solid line Dash dot dot line

Cone/FSS/FSS + Cone/plain patch

0
RHC on cut 1 (at 0°), RHC on cut 1 (at 0°),... (dB)

–10

–20

–30

–40

–50

–60

–70

–80

–90
–180 –150 –120 –90 –60 –30 0 30 60 90 120 150 180
Angle (°)

FIGURE 6.7
Patterns of all four structures compared.
 Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Electric (dB)

FIGURE 6.8
RMS electric field distributions in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Magnetic (dB)

FIGURE 6.9
RMS magnetic field distributions in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.

Circular patch Circular patch with Circular patch with Circular patch with
FSS conical ring conical ring and FSS

–40 –30 –20 –10 0

Power density (dB)

FIGURE 6.10
Time average power density in a slice taken at 7 mm on top of the ground plane at 1.575 GHz.
 RHC pattern
Z

Y
ERHC
Directivity: 9.274 dBi Frequency: 1.575 GHz
Polrzn. loss: –57.604 udB
Material loss: Not known Radial scale: Linear in field magnitude
Mismatch: Not known Ant eff: 100.000% Scale max at: 0 dB–directivity
Gain: 9.274 dBi Rad eff: 100.000% Contour at: –3 dB–directivity

FIGURE 6.11
3-D RHC radiation pattern.

Comparing conical ring, FSS, and ordinary circular patch

Z

ELHC
Directivity: 9.709 dBi Frequency: 1.575 GHz
Polrzn. loss: –60.884 dB
Material loss: –18.581 mdB Radial scale: Linear in field magnitude
Mismatch: Not known Ant eff: 99.573% Scale max at: 0 dB–directivity
Gain: –51.194 dBi Rad eff: 99.573% Contour at: –3 dB–directivity

FIGURE 6.12
3-D LHC radiation pattern.

Antenna under test

Unmanned
aerial vehicle

FIGURE 6.16
A possible antenna test mount for urban canyon environment.
FIGURE 6.17
A KSU Salina UAS asset. (Aerosonde Mark 4.7 UAS platform.)

3° 3°

50 m 50 m

FIGURE 7.3
No obstruction outside 50 m in the bottom center of the antenna as a benchmark over a vertical opening angle of 3°.

Region C

Region A Region B

Runway

FIGURE 7.8
The protected areas of the glide slope.

Flying through two ASs

BGP4

Autonomous system #1 Edge router Autonomous system #2

FIGURE 8.6
ATN/IPS routing infrastructure.
 Propagation losses

Extraneous noise
sources

Atmospheric losses

Uplink
Downlink

Earth station Earth station

FIGURE 9.2
Satellite link budget.

10
Timeout = 30 s
Timeout = 60 s
9
Timeout = 120 s
Timeout = 300 s
8

7
Average throughput (WT/s)

6.1
6 5.93
5.86
5.81
5

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

100
Timeout = 30 s
90 Timeout = 60 s
Timeout = 120 s
Timeout = 300 s
80

70

60
Error WT (%)

50

40

30

20

10 8.01%
4.01%
1.65%
0.41%
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

FIGURE 10.7
Four kinds of throughput and timeout probability curve on the condition of trouble-free equipment.
 10
Timeout = 30 s
9 Timeout = 60 s
Timeout = 120 s
Timeout = 300 s
8
Average throughput (WT/s)

4
3.62
3
2.64
2
1.44
1
0.74

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

100
Timeout = 30 s
Timeout = 60 s
90 Timeout = 120 s
Timeout = 300 s
80

70

60
Error WT (%)

50

40
34.56%
30 29.95%
29.33%
20 20.03%

10

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time ( s )

FIGURE 10.8
Variation curves of 10% of equipment breakdown at the 300th second during the experiment.
 Hadoop server roles

Clients

Distributed data processing Distributed data storage

Map reduce HDFS

Secondary
Job tracker Name node Masters
name node

Data node and Data node and Data node and

task tracker task tracker task tracker
Slaves
Data node and Data node and Data node and
task tracker task tracker task tracker

FIGURE 10.12
Hadoop server roles.

HDFS Split 0 Split 1 Split 2 Split 3

Read Read Read Read

(inputformat) (inputformat) (inputformat) (inputformat)

Mapper Mapper Mapper Mapper

Map Map
task a 1 b 1 c 1 c 1 a 1 c 1 b 1 b 1

Partitioner Partitioner Partitioner Partitioner

Shuffle and sort Shuffle and sort Shuffle and sort

Reduce a 1 a 1 b 1 b 1 b 1 c 1 c 1 c 1 Reduce
task
Reducer Reducer Reducer
Write Write Write
(outputformat) (outputformat) (outputformat)

HDFS Part-0 Part-1 Part-2

FIGURE 10.13
MapReduce data flow.
Aviation

Addresses the Challenges of Modern-Day Air Traffic

Air traffic control (ATC) directs aircraft in the sky and on the ground to safety, while the
Aeronautical Telecommunications Network (ATN) comprises all systems and phases that
assist in aircraft departure and landing. The Aeronautical Telecommunications
Network: Advances, Challenges, and Modeling focuses on the development
of ATN and examines the role of the various systems that link aircraft with the ground.
The book places special emphasis on ATC—introducing the modern ATC system from
the perspective of the user and the developer—and provides a thorough understanding
of the operating mechanism of the ATC system. It discusses the evolution of ATC,
explaining its structure and how it works; includes design examples; and describes all
subsystems of the ATC system.

In addition, the book covers relevant tools, techniques, protocols, and architectures
in ATN, including MIPv6, air traffic control (ATC), security of air traffic management
(ATM), very-high-frequency (VHF) digital link (VDL), aeronautical radio and satellite
communications, electromagnetic interference to aeronautical telecommunications,
quality of service (QoS)-satisfied ATN routing mechanism speed dynamic environments,
and service-oriented architecture (SOA)-based ATN transmission control algorithm.
It also incorporates published research and technical reports to illustrate existing
problems, highlight current methods and opportunities, and consider future directions
and trends.
The authors

• Provide an overview of ATN

• Illustrate the composition of the ATC system
• Explain how to design an ATC system
• Reveal how to use an ATC system to control in-flight airplanes
• Present the results of author research on spatial mitigation
• Introduce the electromagnetic interference effects and response measures
of aviation communications equipment
• Analyze the protective measures of aircraft and ground stations against
electromagnetic interference

The Aeronautical Telecommunications Network: Advances, Challenges,

and Modeling highlights the advances, challenges, and modeling of ATN, and
implements strategies for integrating existing and future data communications networks
into a single internetwork serving the aeronautical industry. This book can aid readers
in working to ensure the effective management of air traffic and airspace, and the
safety of air transport.

K24747
ISBN: 978-1-4987-0504-2
90000

9 781498 705042