CSE 2032

Introduction to
Fog Computing

Module II
 Module II –L1
LO1:OUTLINE THE FEATURES OF COMMUNICATION
MODEL
LO2:DESCRIBE THE ARCHITECTURE OF FOG
COMPUTING
 Introduction
❖ Fog architecture involves services of end devices (switches, routers,
multiplexers, etc) for computational, storage and processing purposes.

❖ Fog computing architecture consists of

1. Physical and logical elements of the network

2. Software, and

3. Hardware to form a complete network of a large number of

interconnecting devices
 Introduction
❖ Fog node distribution (physical as well as geographical, along with the topology
and protocols used form key architectural features of a fog architecture

❖ Fog architecture involves

✔ The distribution of functions at different layers,

✔ The types and the number of protocols used,

✔ The constraints imposed at various layers.

Hierarchical Fog Computing Architecture
 Hierarchical Fog Computing Architecture
Terminal Layer
❖ The terminal layer is the basic layer in fog architecture, includes devices like mobile
phones, sensors, smart vehicles, readers, smartcards, etc.

❖ Devices are distributed across a number of locations separated far apart from each
other.

❖ The layer mostly deals with data sensing and capturing. Devices from different
platforms and different architectures are mainly found in this layer.

❖ The devices have the property of working in a heterogeneous environment, with other
devices from separate technologies and separate modes of communication.
 Hierarchical Fog Computing Architecture
Fog Layer
❖ Fog layer includes devices like routers, gateways, access points, base stations, specific fog servers,
etc., called as Fog nodes.

❖ Fog nodes are located at the edge of a network. An edge can be a hop distance from the end device.

❖ Fog nodes can be static, e.g., located in a bus terminal or coffee shop, or they can be moving, e.g.,
fitted inside in a moving vehicle.

❖ Fog nodes can compute, transfer and store the data temporarily.

❖ Fog nodes and cloud data center connections are enabled by the IP core networks, providing
interaction and cooperation with the cloud for enhancing processing and storage capabilities.
 Hierarchical Fog Computing Architecture
Cloud Layer
❖ This layer consists of devices that can provide large storage and machines (servers) with
high performance.

❖ This layer performs computation analysis and stores data permanently, for back-up and
permanent access to the users.

❖ This layer has high storage and powerful computing capabilities.

❖ Enormous data centers with high computing abilities form a cloud layer.

❖ The cloud layer lies at the extreme end of the overall fog architecture. It acts as a back-up as
well as provides permanent storage for data in a fog architecture
Layered Fog Computing Architecture
 Physical and Virtualization Layer
❖ This layer comprises nodes (Physical and virtual).
❖ The nodes perform the primary task of capturing data and are located at
different locations. Nodes usually involve sensing technology to capture their
surroundings.
❖ Sensors used at this node collect data from the surroundings and collect data
which is then sent to upper layers via gateways for further processing.
❖ A node can be a stand-alone device like a mobile phone or it can be a part of a
large device like a temperature sensor fitted inside a vehicle.
 Monitoring Layer
❖ Nodes can be monitored for the amount of time they work, the temperature
and other physical properties they are possessing, the maximum battery life of
the device, etc.
❖ The performance of applications as well as their present state is also
monitored.
❖ The fog nodes are checked for their energy consumption, the amount of
battery power they consume while performing their tasks.
 Pre Processing Layer
❖ This layer performs various data operations mainly related to analysis.
❖ Data is cleaned and checked for any unwanted data present. Data impurity is
removed and only useful data is collected.
❖ Data analysis at this layer can involve mining meaningful and relevant
information from a vast amount of data collected by the end devices.
❖ Data analysis is one of the essential features that should be taken into
consideration before data is used for a specific purpose.
 Temporary Storage Layer

❖ This layer is associated with non-permanent distribution and replication of

data.
❖ Storage virtualization like VSAN is used in this layer.
❖ Data is removed from the temporary layer once data is moved to the cloud,
from this layer.
 Security Layer

❖ This layer is involved with the privacy of data, the integrity of data, encryption,
and decryption of data.
❖ Privacy in the case of fog computing data can include use-based privacy,
data-based privacy, and location-based privacy.
❖ The security layer ensures secure and preservation of privacy for the data
which is outsourced to the fog nodes.
 Security Layer

❖ This layer is involved with the privacy of data, the integrity of data, encryption,
and decryption of data.
❖ Privacy in the case of fog computing data can include use-based privacy,
data-based privacy, and location-based privacy.
❖ The security layer ensures secure and preservation of privacy for the data
which is outsourced to the fog nodes.
 Mobile Fog Computing (MFC)

❖ Mobile fog computing models are designed to enhance the capabilities

of mobile devices.
❖ It uses fog computing resources, which are located closer to the edge
of the network.
❖ These models aim to improve the performance, efficiency, and user
experience of mobile applications.
 Mobile Fog Computing (MFC)
Offloading Model:
❖ The offloading model involves offloading
Use Cases
computational tasks from mobile devices to fog
nodes or cloud servers.
Image and video
❖ This is done to reduce the computational processing.
burden on the mobile device and take Complex data
advantage of the additional processing power analysis.
available in fog nodes or the cloud.
❖ These models aim to improve the performance,
efficiency, and user experience of mobile
applications.
 Mobile Fog Computing (MFC)
Augmentation Model:
❖ The augmentation model extends the
Use Cases
capabilities of mobile devices by providing
them with additional resources, such as
computational power, storage, or specialized
sensors, trough fog nodes..
Real-time language translation.
High-quality gaming on mobile devices.
Image recognition with the help of specialized
cameras.
 Mobile Fog Computing (MFC)
Delegation Model:
❖ The delegation model involves dividing a task into subtasks and
distributing them between mobile devices and fog nodes for parallel
processing.
❖ It optimizes the use of resources by delegating tasks based on device
capabilities and proximity to data sources.
Collaborative applications where multiple users work on a shared
task.
Distributed sensing and monitoring in IoT applications.
 Fog Architecture for Smart Cities
❖ The new challenges of big data analysis posed by smart cities
demands to develop novel and high-performance computing
architectures
❖ The rising of Cloud Computing and Cloud Storage in industry provides
a solution to support dynamic scalability in many smart city
applications.
✔Smart house
✔Smart lighting
✔Video surveillance
✔Business and academic computing tasks in education institutions
 Fog Architecture for Smart Cities
❖ However, the deployment of massive numbers of sensors in future
smart cities requires
An event is a change in state, or an update, like
✔ location awareness an item being placed in a shopping cart on an
e-commerce website
✔ low latency,
❖ which are lacking in current commercial Cloud Computing models
❖ Rather than analyzing all original data from massive sensor networks,
an event-driven platform can greatly reduce the computational burden
of smart cities. Smart house
Fog Architecture for Smart Cities
 Fog Architecture for Smart Cities
Layer 4
❖ At the edge of network, layer 4, is the sensing network which contains
numerous sensory nodes.
❖ Those sensors are noninvasive, highly reliable, and low cost.
❖ They can be widely distributed at various public infrastructures to
monitor their condition changes over time.
❖ Note that massive sensing data streams are generated from these
sensors that are geospatially distributed, which have to be processed
as a whole.
 Fog Architecture for Smart Cities
Layer 3
❖ The nodes at the edge forward the raw data into the next layer, layer 3.
❖ It is comprised of many low-power and high-performance computing
nodes or edge devices.
❖ Each edge de-vice is connected to and responsible for a local group of
sensors that usually cover a neighborhood or a small community,
performing data analysis in a timely manner
 Fog Architecture for Smart Cities
Layer 3
❖ The output of the edge device has two parts:
❖ 1. Reports of the results of data processing to an intermediate
computing node at its next upper layer,
❖ 2. Simple and quick feedback control to a local infrastructure to
respond to isolated and small threats to the monitored infrastructure
components
 Fog Architecture for Smart Cities
Layer 2
❖ Layer 2 consists of a number of intermediate computing nodes, each
of which is connected to a group of edge devices at layer 3 and
associates spatial and temporal data to identify potential hazardous
events.
❖ Meanwhile, it makes quick response to control the infrastructure when
hazardous events are detected.
❖ The quick feedback control provided at layers 2 and 3 acts as localized
“reflex” decisions to avoid potential damage.
 Fog Architecture for Smart Cities
Layer 2
❖ Example:
❖ if one segment of gas pipeline is experiencing a leak-age or a fire is
detected, these computing nodes will detect the threat and shutdown
the gas supply to this section.
❖ Meanwhile, all the data analysis results at these two layers are
reported to the top layer, for more complex and large-scaled behavior
analysis and condition monitoring.
 Fog Architecture for Smart Cities
Layer 1
❖ The top layer is a Cloud Computing data center, providing city-wide
monitoring and centralized controlling.
❖ The high-performance distributed computing and storage capacity
allows us to perform complex, long-term (days to years), and city-wide
behavior analysis at this layer, such as
This allows municipalities to perform city-wide
✔ large-scale event detection, response and resource management in the case of a
natural disaster or a large-scale service interruption
✔ long-term pattern recognition, and
✔ Relationship modeling, to support dynamic decision making.
Fog Architecture for Health Care
 Introduction to IEEE 802.11
❖ IEEE 802.11 is part of the IEEE 802 set of local area
network (LAN) technical standards.
❖ It specifies the set of media access control (MAC) and physical
layer (PHY) protocols for implementing wireless local area
network (WLAN) computer communication
❖ IEEE 802.11 is used in most home and office networks to allow
laptops, printers, smartphones, and other devices to communicate
with each other and access the Internet without connecting wires.
 Introduction to IEEE 802.11
• Each computer, mobile, portable or fixed, is referred to as a station in
802.11.
• The difference between a portable and mobile station is that a portable
station moves from point to point but is only used at a fixed point.
• Mobile stations access the LAN during movement. Fundamental to the
IEEE 802.11 architecture is the concept of a Basic Service Set (BSS) or
wireless LAN cell.
• A BSS is defined as a group of stations that coordinate their access to
the medium under a given instance of medium access control.
• The geographic area covered by a BSS is known as the Basic Service
Area (BSA), which is very similar to a cell in a cellular communication
network.
• All stations within a BSA with tens of meters in diameter may
 IEEE 802.11 Architecture
• The 802.11 standard supports the formation of two distinct types of
BSSs:
1) ad hoc network and
2) Infrastructure BSS.
• Two or more BSSs are interconnected using a Distribution System or
DS.
 IEEE 802.11 Architecture
• The first type of BSS is known as an ad
hoc network, which consists of a group
of stations within the range of each
other.
• As its name implies, ad hoc networks are
temporary in nature, which are typically
created and maintained as needed
without prior administrative arrangement.
• Ad hoc networks can be formed
anywhere spontaneously and can be
disbanded after a limited period of time
 IEEE 802.11 Architecture
• The second type of BSS is known as
infrastructure BSS (IBSS), which is commonly
used in practice. Here, several BSSs are
interconnected by a distribution system to form
an extended service set(ESS)
• The BSSs are like cells in a cellular
communications network.
• Each BSS is provided with an Access point (AP)
that has station functionality and provides
access to the distribution system.
• APs are located such that the BSSs they serve
overlap slightly to provide continuous service to
 IEEE 802.11 Architecture
• If station A associated with AP-1 wants to send a frame
to another station associated with AP- 2, the first sends
a frame to its access point (AP-1), which forwards the
frame across the distribution system to the access
point AP-2. AP-2 finally delivers it to the destination
station.
• The technique used for this purpose is known as
scanning, which involves the following steps:
1. A station sends a probe frame.
2. All APs within reach reply with a probe response
frame.
3. The station selects one of the access points, and
sends the AP an Association Request frame.
4. The AP replies with an Association Response
 WPAN Short Range Technologies
• Wireless Personal Area Networks (WPAN) cover a relatively small area,
typically within the range of 10 meters in radius.
• The development of WPAN can be traced back to the early days of
computing when cables were used to connect devices like printers,
modems, and keyboards.
• WPAN technologies are designed for personal and local area
networking, and they are often used for connecting devices like
smartphones, tablets, laptops, and various other peripherals and
sensors.
• The first commercial products with built-in Bluetooth capabilities was
Motorola Timeport P7689 mobile phone with an embedded Bluetooth
module and an IBM ThinkPad A30 laptop computer with an optional PC
 WPAN Short Range Technologies
• Another critical development in WPAN technology was IEEE 802.15
group.
• This group aimed at creating standards for low-power wireless
communications between nearby devices that could operate without
interfering with each other or other Wi-Fi networks operating nearby.
• One of the most successful implementations of the IEEE 802.15
standard was ZigBee – a low-cost mesh networking protocol widely
used for home automation systems, including
o lighting control,
o heating/cooling systems control, and
o security monitoring solutions.
 WPAN Short Range Technologies
As per IEEE, Wireless Personal Area Network is classified into 3 classes
i.e
1.High-rate WPAN (HR-WPAN) : It is defined in the IEEE 802.15.3 standard.
Data throughput is > 20 Mbps.
2.Medium-rate WPAN (MR-WPAN) : It is defined in the IEEE 802.15.1
standard. Data throughput is 1 Mbp.
3.Low-rate WPAN (LR-WPAN) : It is defined in the IEEE 802.15.4 standard.
Data throughput is < 0.25 Mbps.
 WPAN Short Range Technologies
As per IEEE, Wireless Personal Area Network is classified into 3 classes
i.e
1.High-rate WPAN (HR-WPAN) : It is defined in the IEEE 802.15.3 standard.
Data throughput is > 20 Mbps.
2.Medium-rate WPAN (MR-WPAN) : It is defined in the IEEE 802.15.1
standard. Data throughput is 1 Mbp.
3.Low-rate WPAN (LR-WPAN) : It is defined in the IEEE 802.15.4 standard.
Data throughput is < 0.25 Mbps.
WPAN Short Range Technologies
WPAN Short Range Technologies
 LPWAN and other mediums
❑ LPWANs (Low-Power Wide-Area Networks) are a type of radio-based
connectivity used for wireless data communication for IoT devices and
M2M (machine-to-machine) applications.
❑ It is a class of network technologies designed to communicate
wirelessly over relatively long distances using lower power than in
case of other networks, such as telephony, satellite communications or
Wi-Fi
Energy-efficiency and Needs of ever-evolving
Wide signal coverage market of IoT applications.
 LPWAN and other mediums
❑ LPWANs (Low-Power Wide-Area Networks) are a type of radio-based
connectivity used for wireless data communication for IoT devices and
M2M (machine-to-machine) applications.
❑ It is a class of network technologies designed to communicate
wirelessly over relatively long distances using lower power than in
case of other networks, such as telephony, satellite communications or
Wi-Fi
Energy-efficiency and Needs of ever-evolving
Wide signal coverage market of IoT applications.
 LPWAN and other mediums
LTE-M (CAT-M) LONG TIME EVOLUTION –
MACHINE TYPE COMMUNICATION

•LTE-M is a leading LPWA network

technology for Internet of
Things applications.
•Used for interconnecting objects such
as IoT sensors and actuators, or other
industrial devices via radio modules.
LPWAN and other mediums
 LPWAN and other mediums
LTE-M (CAT-M)

• It is a standard for narrow-bandwidth cellular communications to

connect resource-constrained devices to the Internet that transmit
small amounts of data over long periods of time.
• High data throughput rates that enable voice and video transmission
as well as faster upgrades.
• Support for moving assets so mobile devices can function without
interruption.
• Lower data transmission costs.
 LPWAN and other mediums
NB-IoT (Narrowband Internet of Things)

• NB-IoT is another LPWAN technology designed to reduce power consumption

and improve system capacity and spectrum efficiency, especially in deep and
remote locations.
• Lower device manufacturing costs derived from simpler construction and
leveraging chipsets specifically engineered for NB-IoT.
• Power efficiency due to its simpler waveform and the ability to “sleep” when not
in use.
• Battery life of more than 10 years can be supported for a wide range of use cases.
• Security (user privacy, authentication, and data integrity) by leveraging the
NB-IoT (Narrowband Internet of Things)
NB-IoT (Narrowband Internet of Things)
 Long Range Technologies
• Long range IoT wireless technologies form the basis for a LPWAN (Low
Power Wide Area Network).
• In these types of networks, end devices with low energy consumption –
typically sensors – are connected to gateways which transmit data to other
devices and network servers.
• The network devices assess the received data and control the end devices.
• Accordingly, the protocols are specially designed for long-range
capabilities, low-power devices, and reduced operating costs.
 Long Range Technologies
SigFox
• SigFox is a tailor-made solution for long ranges (30-50km in rural areas,
3-10km in urban areas), low data rates (12 bytes per message, max. 140
messages a day per end device), and preferably low power operation.
• End devices equipped with SigFox technology transmit data to SigFox base
stations which then forward the data to SigFox servers.
• This is where the data are processed before the results are sent back to the
respective end devices for visualization.
• This means: Data are managed by SigFox on in its own cloud servers.
 Long Range Technologies

Sigfox offers end-to-end connectivity through LPWAN, which sends small

data packets (100bps) from sensors, via the Sigfox network and Cloud to
connect sensors to digital third party sites and applications
 Long Range Technologies
LoRa
• LoRa is a wireless technology that offers long range, low power and secure data
transmission for M2M and IoT applications
• LoRa can be used to connect sensors, gateways, machines, devices, animals,
people etc. wirelessly to the cloud.
• LoRa signals have excellent penetration capabilities, allowing them to pass through
obstacles like walls and equipment. This makes LoRa technology suitable for indoor
and outdoor applications.
• LoRaWAN defines the communication protocol and network architecture, enabling
the efficient and secure transmission of data from LoRa devices to network servers
and applications..
 Long Range Technologies

LoRa is the physical layer i.e. the chip and LoRaWAN is the MAC layer i.e. the software that’s put
on the chip to enable networking.
 5G Standards
❖ 5G architecture is needed for several reasons, as it addresses the limitations
of previous generations of wireless technology (such as 4G, 3G, and 2G) and
provides significant benefits to both consumers and industries.
1. Higher Data Rates: Higher data rates compared to previous generations
(2Mbps/20Mbps/20 Gbps).
2. Lower Latency: Ultra-low latency, reducing the delay in data
transmission(Fog Computing / Network Slicing).
5G Standards
3. Massive Connectivity: Can handle a massive number of
connected devices, making it suitable for the Internet of
Things (IoT).
4. Network Slicing: Creation of virtual networks tailored to
specific use cases or services/ Same physical infrastructure
can be used to support a wide range of applications with
varying requirements, from high bandwidth to low latency.
5. Enhanced Mobile Broadband: superior mobile broadband
experience
5G Protocol Stack
 5G Architecture
OWA Layer: OWA layer is the short form of Open Wireless Architecture
layer. It functions as physical layer and data link layer of OSI stack.
Network Layer: It is used to route data from source IP device to the
destination IP device/system. It is divided into lower and upper network
layers
Open Transport Layer: It combines functionality of both transport layer
and session layer.
Application Layer: It marks the data as per proper format required. It also
does encryption and decryption of the data. It selects the best wireless
connection for given service.
5G Architecture
 5G Architecture
1) User Equipment (UE):
• This is the endpoint device.
• Smartphone, tablet, IoT device, or any device capable of connecting to a 5G
network.
• UEs are responsible for transmitting and receiving data to and from the
network.
1) Radio Access Network (RAN):
• It connects the UE to the core network.
• In 5G, the RAN is more distributed and utilizes various technologies
• Provide efficient and high-capacity connections.
5G Architecture
• RAN includes:
• New Radio (NR) - air interface technology
• Small Cells - use small cells to provide localized coverage and capacity
/ urban areas, stadiums, and other high-density environments to
enhance the user experience.
• Dynamic Spectrum Sharing (DSS): DSS enables the efficient use of
spectrum by allowing both 4G and 5G to coexist on the same frequency
bands, making the transition to 5G smoother.
5G Architecture
• Core Network (CN): The core network is the central part of the 5G
architecture. It is responsible for managing and routing traffic and supporting
various network functions, including:
1. Network Function Virtualization (NFV): 5G core networks use
virtualization technologies, allowing network functions to run on virtual
machines rather than dedicated hardware.
2. Software-Defined Networking (SDN): SDN enhances network
management by separating the control plane from the data plane, allowing
for dynamic and programmable network configurations.
5G Architecture
• Network Function Virtualization (NFV):
• NFV allows for the separation of communication services from
dedicated hardware, such as routers and firewalls.
• This separation means network operations can provide new services
dynamically and without installing new hardware.
• Deploying network components with network functions virtualization
takes hours instead of months like with traditional networking
5G Architecture
• Software-Defined Networking (SDN):
is responsible for providing network services and
applications to end-users.

It consists of a centralized controller that

communicates with the infrastructure layer devices
and is responsible for managing and configuring the
network.

It consists of physical and virtual network devices

such as switches, routers, and firewalls that are
responsible for forwarding network traffic based on
the instructions received from the control plane.
 5G Architecture
• Network Slicing
• 5G networks support network slicing,
which allows for the creation of multiple
virtual networks on a common physical
infrastructure.
• Each slice can be customized to meet
specific requirements, such as low
latency for autonomous vehicles, high
bandwidth for video streaming, or low
power for IoT devices.
 5G Architecture
• Multi-Access Edge Computing (MEC):
• MEC brings computing resources closer
to the network edge.
• This reduces latency by processing data
locally, making it suitable for
latency-sensitive applications like
augmented reality, virtual reality, and IoT.
 5G Architecture
• The Authentication Server Function (AUSF) allows to authenticate the UE
and access services of the 5G core.
• The Session Management Function (SMF) is the control function that
manages the user sessions including establishment, modification and
release of sessions, and it can allocate IP addresses.
• Policy Control Function (PCF): Policy rules for application and service data
flow detection, QoS, and flow based charging to the SMF.
• Application Function (AF): It plays a key role in traffic management and
QoS assignments, through interaction with the policy elements.
 Stateless and Stateful Protocols
❑ It is a network model in which the client sends a request to the server
and the server in return sends a response back according to the current
state just like the Request-Response model.
❑ The server does not keep the session information or the status of each
communication partner for multiple requests.
❑ They are very easy to implement on the Internet.
❑ Stateless protocols work better when the crash occurs because no state
needs to be restored, a failed server can simply reboot after a crash.
❑ Examples:- HTTP (Hypertext Transfer Protocol), UDP (User Datagram
Protocol), DNS (Domain Name System).
Stateless and Stateful Protocols
In this protocol, suppose a client sends a request to the server
and the server doesn't respond, then the client resends a
request to the server.
Stateful protocols are logically heavy to implement on the
Internet.
Stateful Protocol does not work better at the time of the crash
because stateful servers must retain information about the
state and session details of internal states.
Examples:- FTP (File Transfer Protocol), Telnet.
Stateless and Stateful Protocols
 Test I [ Answer All Questions]
1. What is the need of Fog Computing. [3]
2. Classify types of IP addressing [3]

3. List any three difference between Cloud, Fog and Edge Computing[4]

4. Write down the characteristics of Fog Computing [5]

5. Explain Fog Deployment Models in detail (10)