Characterising Network Traffic

6CS029 Advanced Networking

The Need to Scale the Network
• Organizations increasingly rely on their network
infrastructure to provide mission-critical services.
• Evolving organizations require networks that can scale
and support:
– Converged network traffic
– Critical applications
– Diverse business needs
– Centralized administrative control
• Network designs include small networks that use a single
LAN switch, up to very large networks with thousands of
connections.
The Need to Scale the Network
 Existing Network

What’s the Starting Point?

“If we could first know where we are

and whither we are tending, we
could better judge what to do and
how to do it.”
Abraham Lincoln
Existing Network
How do we evaluate the
health of the existing
network?
• Performance
• Availability
• Bandwidth utilization
• Accuracy
• Efficiency
• Response time
• Status of major routers,
switches, firewalls, servers
Characterising Network Traffic
• Understand how network users interact
with network resources and services
• Identify all user groups
• Assess the impact of adding new user
groups
Characterising Network Traffic
What are the most important traffic factors to
be taken into consideration?
• Traffic flow
• Location of traffic sources and data stores
• Traffic load
• Traffic behaviour
• Quality of Service (QoS)
requirements
User Communities
UserCommunity Size of Location(s)of Application(s)
Name Community Community Used by
(Number of Community
Users)
Data Stores
Data Store Location Applications Used by User
Communiti
es
Traffic Flow
Destination 1 Destination 2 Destination 3 Destination n
MB/sec MB/sec MB/sec MB/sec

Source 1

Source 2

Source 3

Source n
Types of Traffic Flow

• Terminal/host
• Client/server
• Thin client
• Peer-to-peer
• Server/server
• Distributed computing
QoS Requirements
What makes traffic flow in voice over IP networks
challenging to characterise and plan for?
• The flow associated with transmitting the audio
voice is separate from the flows associated
with call setup and teardown.
– The flow for transmitting the digital voice is
essentially peer-to-peer.
– Call setup and teardown is a client/server flow
• QoS requirements for multimedia traffic
Network Applications
Traffic Characteristics
Name of Type of Protocols User Data Approximate QoS
Application Traffic Used by Communities Stores Bandwidth Requirements
Flow Application That Use the (Servers, Requirements
Application Hosts, etc)
Traffic Load
• Required capacity
– The number of stations
– The average time that a station is idle between
sending frames
– The time required to transmit a message once
medium access is gained
• Traffic behaviour
– Broadcasts
– Multicasts
Typical Bottlenecks?
• Shared services (centralised servers etc)
• Multi-user applications and databases
• Low-speed NICs
• Shared LAN segments
• Low-bandwidth WAN links
• Core routing and switching components
• Firewalls (particularly public-facing)
• Inappropriate compression usage
 Green Inc. is expanding and just
got another floor on the building
they have their main office. The
new floor will also need to be
connected to Green’s network.

Within Green Inc. Network, network 1 users

need to frequently access a database stored in
Server 1.
Network 2 users also frequently access a
database stored in Server 2.

Task1: Identify performance, scalability and redundancy problems with

this network design.
 Green Inc. is expanding and just
got another floor on the building
they have their main office. The
new floor will also need to be
connected to Green’s network.

Within Green Inc. Network, network 1 users

need to frequently access a database stored in
Server 1.
Network 2 users also frequently access a
database stored in Server 2.

Task2: Propose a new network design to solve the problems

identified
Borderless Switched Networks

• A network architecture that can connect anyone,

anywhere, anytime, on any device; securely,
reliably, and seamlessly.
• It provides the framework to unify wired and
wireless access, built on a hierarchical
infrastructure of hardware that is scalable and
resilient.
• Borderless switched networks are hierarchical,
modular, resilient, and flexible.
Hierarchy in the Borderless Switched
Network
Hierarchical networks use a tiered design of access,
distribution, and core layers with each layer performing a
well-defined role in the campus network.
Three-tier layer Two-tier layer
Access, Distribution, and Core Layer
Functions
Access Layer
• The access layer provide network access to the user.
• Access layer switches connect to distribution layer switches.
Distribution Layer
• The distribution layer implements routing, quality of service, and security.
• It aggregates large-scale wiring closet networks and limits Layer 2 broadcast
domains.
• Distribution layer switches connect to access layer and core layer switches.
Core Layer
• The core layer is the network backbone and connects several layers of the
network.
• The core layer provides fault isolation and high-speed backbone connectivity.
Advantages of Hierarchical
Network Design

• Scalability
• Easier to add to the network
• Manageability
• Easier to identify location of problems
• Broadcast traffic segmentation
• Traffic confined to smaller broadcast
domains
• Less traffic over expensive links
Three-Tier and Two-Tier
Examples
Three-tier Campus Network
• Used by organizations requiring access,
distribution, and core layers.
• The recommendation is to build an extended-star
physical network topology from a centralized
building location to all other buildings on the same
campus.

Two-tier Campus Network

• Used when separate distribution and core layers
is not required.
• Useful for smaller campus locations, or in campus
sites consisting of a single building.
• Also known as the collapsed core network design.
Design for Scalability
Scalability is the term for a network that can grow without
losing availability and reliability.

Network designers must develop strategies to enable the

network to be available and to scale effectively and easily.

This is accomplished using:

• Redundancy
• Multiple Links
• Scalable Routing protocol
• Wireless Connectivity
Plan for Redundancy
Redundancy can prevent disruption of network services by minimizing the
possibility of a single point of failure by:
• Installing duplicate equipment
• Providing failover services for critical devices

Redundant paths offer alternate

physical paths for data to traverse the
network supporting high availability.
• However, redundant paths in an
Ethernet network may cause
logical Layer 2 loops.
• Therefore, Spanning Tree Protocol
(STP) is required.
Reduce Failure Domain Size
Reduce Failure Domain Size
A well-designed network controls traffic and limits the size
of failure domains (i.e., the area of a network that is
impacted when the network experiences problems).
• In the hierarchical design model, failure domains are terminated at
the distribution layer.
• Every router functions as a gateway for a limited number of access
layer users.
Routers, or multilayer switches, are usually deployed in
pairs in a configuration referred to as a building, or
departmental, switch block.
• Each switch block acts independently of the others.
• As a result, the failure of a single device does not cause the network
to go down.
Increase Bandwidth
Link aggregation (e.g., EtherChannel) allows an administrator to
increase the amount of bandwidth between devices by creating
one logical link made up of several physical links.

• EtherChannel combines existing

switch ports into one logical link
using a Port Channel interface.
• Most configuration tasks are done on
the Port Channel interface (instead
of on each individual port) to ensure
configuration consistency on the
links.
• EtherChannel can load balance
between links.
Expand the Access Layer
An increasingly popular option for extending access layer connectivity is
through wireless.
• Wireless LANs (WLANs) provides increased flexibility, reduced costs,
and the ability to grow and adapt to changing network and business
requirements.

• To communicate wirelessly, end devices

require a wireless NIC to connect to a wireless
router or a wireless access point (AP).
Considerations when implementing a wireless
network include:
• Types of wireless devices connecting to the
WLAN
• Wireless coverage requirements
• Interference considerations
• Security considerations
Tune Routing Protocols
Advanced routing protocols, such as Open Shortest Path
First (OSPF) are used in large networks.
• OSPF is a link-state routing
protocol that uses areas to
support hierarchical networks.
• OSPF routers establish and
maintain neighbor adjacencies
with other connected OSPF
routers.
• OSPF routers synchronize their
link-state database.
• When a network change
occurs, link-state updates are
sent, informing other OSPF
routers of the change and
establishing a new best path, if
one is available.
Multilayer Switching
Multilayer switches are typically deployed in the core and
distribution layers of an organization's switched network.
• They support some routing protocols and forward IP
packets at a rate close to that of Layer 2 forwarding.
• Multilayer switches often support specialized hardware,
such as application-specific integrated circuits (ASICs).
• ASICs along with dedicated software data structures can
streamline the forwarding of IP packets independent of
the CPU.
VLANs

• Logical hierarchy imposed on a flat

switched network allowing
• Scalability
• Formation of workgroups
• Simplified admin
• Better security
Summary

• Understand network traffic in terms of:

• Flow
• Load
• Behavior
• QoS requirements
Questions?