Routing in IP-based Networks

Abstract

Routing Protocols are very important in the modern day’s communications, as
with the increasing number of network devices it has become very vital to understand
and implement the correct routing protocol for the network and to process and
communicate the huge routing information among all the network nodes. Dynamic
Routing protocol selects the best optimal path from source node to destination node.
There are three major types of dynamic routing protocols, each having their own
criteria for the route selection and propagation. These are categorized as.
1) Distance Vector
2) Linked Stated
3) Hybrid
The major example ofaDistance Vector Routing (DVR) Protocol isaRouting
Information Protocol (RIP). RIPais normally used for smaller networks and its hop
limit is 15. On the other hand, two other routing protocols, i.e., EnhancedaInterior
Gateway Routing Protocola (EIGRP) and Open Shortest PathaFirst (OSPF) are
designed for the huge networks and are much more efficient routing protocols as
compared toaDistance Vector Routing Protocols. EIGRP falls under the category of
Hybrid Routing Protocols or sometime it’s called as advanced distance vector routing
protocol. EIGRP is Cisco proprietary routing protocol. Open Shortest a Path First
(OSPF) is widely used a routing protocol based on Dijkstra’s Algorithm. OSPF is open
standard routing protocol as it is not bound to any specific vendor equipment. All
these routing protocols have different route selection mechanism, different
architecture, route processing, convergence and delay. In this thesis we present a
study of these routing protocols for the real time application using OPNET. In order
to study the performance of RIP, EIGRP and OSPF, we present different scenarios to
evaluate which routing protocols efficiency is better than the others. The performance
evaluation of these routing protocols is based on different metrics like convergence time,
end-to-end delay, throughput, jitter, and packet loss. Results show that EIGRP routing
protocol provides a better performance than OSPF and RIP routing protocols for real time
applications.

Full text

based NetworksRouting in IP-
By
Mr. Murad Khan
CIIT/SP11-REE-047/ISB
MS Thesis
In
Electrical Engineering
COMSATS Institute of Information Technology
Islamabad – Pakistan
Spring, 2012

Routing in IP-based Networks
A Thesis presented to
COMSATS Institute of Information Technology
In partial fulfillment
of the requirement for the degree of
MS (Electrical Engineering)
By
Mr. Murad Khan
CIIT/SP11-REE-047/ISB
Spring, 2012
COMSATS Institute of Information Technology
ii

Engineering).
A post Graduate Thesis submitted to Department of Electrical Engineering as
partial fulfillment of the requirement for the award of Degree of M.S
(Electrical
Supervisor:
Dr. Nadeem Javaid,
Assistant Professor,
Department of Electrical Engineering,
COMSATS Institute of Information Technology (CIIT)
Islamabad Campus
June, 2012
Name Registeration Number
Mr. Murad Khan CIIT/SP11-REE-047/ISB
Routing in IP-based Networks
iii

Final Approval
This thesis titled
Routing in IP-based Networks
By
Mr. Murad Khan
CIIT/SP11-REE-047/ISB
Hasabeenaapproved
For the COMSATSaInstitute of Information Technology, Islamabad
External Examiner: __________________________________
Supervisor: ________________________
Dr. Nadeem Javaid /Assistant professor
Department of Electrical Engineering
Islamabad Campus
HoD: ________________________
Dr. Shafayat Abrar / Associate professor
Department of Electrical Engineering
Islamabad Campus
iv

Declaration
I Mr. Murad Khan, CIIT/SP11-REE-047/ISB hereby declare that I have produced
the work presented in this thesis, during the scheduled period of study. I also declare
that I have not taken any material from any source except referred to wherever due
that amount of plagiarism is within acceptable range. If a violation of HEC rules on
research has occurred in this thesis, I shall be liable to punishable action under the
plagiarism rules of the HEC.
Date: ________________
________________
Mr. Murad Khan
CIIT/SP11-REE-047/ISB
v

Certificate
It is certified that Mr. Murad Khan, CIIT/SP11-REE-047/ISB hasacarried
outaall the work relatedato thisathesis underamyasupervision at the Departmentaof
Electrical EngineeringaCOMSATS Institute of InformationaTechnology, Islamabad
and the work fulfillsathe requirements foraaward of MS degree.
Date: _________________
Supervisor:____________________
Dr. Nadeem Javaid /Assistant professor
Department of Electrical Engineering
CIIT Islamabad Campus
Head of Department:
____________________________
Dr. Shafayat Abrar/Associate professor
HoD Electrical Engineering
vi

DEDICATION
Dedicated to my family and friends
vii

ACKNOWLEDGMENT
I amaheartily grateful toamy supervisor, Dr. Nadeem Javed, whoseapatient
encouragement, guidanceaand insightful criticismafrom the beginning to theafinal
levelaenabled me have aadeep understandingaof theathesis.
Lastly, I offer my profound regard and blessing to everyone who supported me in
any respect during the completion of my thesis.
Mr. Murad Khan
CIIT/SP11-REE-047/ISB
viii

ABSTRACT
Routing Protocols are very important in the modern day’s communications, as
with the increasing number of network devices it has become very vital to understand
and implement the correct routing protocol for the network and to process and
communicate the huge routing information among all the network nodes. Dynamic
Routing protocol selects the best optimal path from source node to destination node.
There are three major types of dynamic routing protocols, each having their own
criteria for the route selection and propagation. These are categorized as.
1) Distance Vector
2) Linked Stated
3) Hybrid
The major example ofaDistance Vector Routing (DVR) Protocol isaRouting
Information Protocol (RIP). RIPais normally used for smaller networks and its hop
limit is 15. On the other hand, two other routing protocols, i.e., EnhancedaInterior
Gateway Routing Protocola(EIGRP) and Open Shortest PathaFirst (OSPF) are
designed for the huge networks and are much more efficient routing protocols as
compared toaDistance Vector Routing Protocols. EIGRP falls under the category of
Hybrid Routing Protocols or sometime it’s called as advanced distance vector routing
protocol. EIGRP is Cisco proprietary routing protocol. Open ShortestaPath First
(OSPF) is widely usedarouting protocol based on Dijkstra’s Algorithm. OSPF is open
standard routing protocol as it is not bound to any specific vendor equipment. All
these routing protocols have different route selection mechanism, different
architecture, route processing, convergence and delay. In this thesis we present a
study of these routing protocols for the real time application using OPNET. In order
to study the performance of RIP, EIGRP and OSPF, we present different scenarios to
evaluate which routing protocols efficiency is better than the others. The performance
evaluation of these routing protocols is based on different metrics like convergenceotime,
end-to-end delay, throughput, jitter, andopacket loss. Results showothat EIGRP routing
protocol providesabetter performance thanaOSPF and RIP routing protocols for real time
applications.
ix

Table of Contents
1 Introduction 2
1.1 Routingprotocols............................ 2
1.2 Motivation................................ 4
1.3 ProblemStatement........................... 4
1.4 Thesisorganization........................... 5
2 Routing Protocols - Background 7
2.1 Overview................................. 7
2.2 Routing Protocols Attribute and Characteristics . . . . . . . . . . . 8
2.2.1 Best Possible Routes . . . . . . . . . . . . . . . . . . . . . . 8
2.2.2 Faster Convergence . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.3 SecurityOptions ........................ 8
2.2.4 AvoidingLoops......................... 8
2.3 Important Concepts About Routing and Metrics . . . . . . . . . . . 8
2.3.1 Metrics ............................. 9
2.3.2 NeedofaMetric ........................ 9
2.3.3 Parameters of a Metric . . . . . . . . . . . . . . . . . . . . . 9
2.4 classification of Routing Protocols . . . . . . . . . . . . . . . . . . 10
2.5 Static and Dynamic Routing . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Classfull Routing Protocols and Classless Routing Protocols . . . . 11
2.6.1 Classfull Routing . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6.2 Classless Routing . . . . . . . . . . . . . . . . . . . . . . . . 12
2.7 Linked-State Routing . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.7.1 Link-state protocol Characteristics . . . . . . . . . . . . . . 15
2.7.2 Methods of Routing . . . . . . . . . . . . . . . . . . . . . . 15
2.7.3 Strength and Weaknesses of LSR . . . . . . . . . . . . . . . 15
2.8 Distance-Vector Routing . . . . . . . . . . . . . . . . . . . . . . . . 16
2.8.1 Characteristics of Distance Vector Routing . . . . . . . . . . 17
2.8.2 Pros and Cons of DVR . . . . . . . . . . . . . . . . . . . . . 17
x

3 Dynamic Routing Protocols 19
3.1 Routing Information Protocol . . . . . . . . . . . . . . . . . . . . . 19
3.1.1 SplitHorizon .......................... 22
3.1.2 Route Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1.3 PoisonReverse ......................... 22
3.1.4 Hold Down Timers . . . . . . . . . . . . . . . . . . . . . . . 24
3.1.5 Triggered Updates . . . . . . . . . . . . . . . . . . . . . . . 24
3.1.6 Counting to Infinity . . . . . . . . . . . . . . . . . . . . . . 25
3.1.7 Comparison ........................... 25
3.2 Enhanced Interior Gateway Routing Protocol (EIGRP) . . . . . . . 26
3.2.1 Neighbor Tables . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.2 TopologyTable ......................... 26
3.2.3 RoutingTable.......................... 27
3.2.4 EIGRP Features . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 Open Shortest Path First (OSPF) . . . . . . . . . . . . . . . . . . . 27
3.3.1 RoutingTable.......................... 28
3.3.2 TopologyTable ......................... 28
3.3.3 NeighborTable:......................... 29
3.3.4 OSPF Area Design and Principles . . . . . . . . . . . . . . . 29
3.3.5 OSPF Neighbor Formation . . . . . . . . . . . . . . . . . . . 29
3.3.6 OSPF Packet Types . . . . . . . . . . . . . . . . . . . . . . 31
3.3.6.1 Types of LSAs . . . . . . . . . . . . . . . . . . . . 31
3.3.7 OSPFCost ........................... 32
3.3.8 Understanding DR and BDR . . . . . . . . . . . . . . . . . 33
3.3.9 OSPF Virtual Link . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.10 OSPF Summarization . . . . . . . . . . . . . . . . . . . . . 33
3.3.11 OSPF Route Filtering . . . . . . . . . . . . . . . . . . . . . 35
3.3.12 OSPF Area Types . . . . . . . . . . . . . . . . . . . . . . . 35
4 Simulation Results 37
4.1 Simulation Environment . . . . . . . . . . . . . . . . . . . . . . . . 37
4.1.1 OPNET Structure . . . . . . . . . . . . . . . . . . . . . . . 37
4.1.1.1 Hierarchical Structure . . . . . . . . . . . . . . . . 37
4.1.2 How to Analyze and Design in OPNET . . . . . . . . . . . . 39
4.1.3 OPNET Environment . . . . . . . . . . . . . . . . . . . . . 39
4.2 SimulationResults ........................... 41
5 Conclusions 48
References 50
xi

List of Figures
1.1 Distance vector protocol . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 Classification of Routing Protocols . . . . . . . . . . . . . . . . . . 10
2.2 Network Deployed with Classful Routing and Same Subnet Mask . 12
2.3 Network Deployed with Classless Routing Protocol . . . . . . . . . 13
2.4 Link State Data Structure . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Link State Routing [5] . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Distance Vector Routing [5] . . . . . . . . . . . . . . . . . . . . . . 16
3.1 RIP-Ver1 Packet Format [1] . . . . . . . . . . . . . . . . . . . . . . 19
3.2 RIP-Ver2 Packet Format [1] . . . . . . . . . . . . . . . . . . . . . . 20
3.3 RIPng Packet Format [1] . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Routing Loops in RIP [1] . . . . . . . . . . . . . . . . . . . . . . . . 21
3.5 Example of Split Horizon [1] . . . . . . . . . . . . . . . . . . . . . . 22
3.6 Example of Route Poisoning [1] . . . . . . . . . . . . . . . . . . . . 23
3.7 PoisonReverse[1]............................ 23
3.8 HoldDownTimers[1] ......................... 24
3.9 Triggered Updates [1] . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.10 Counting to Infinity [1] . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.11 Example of Counting to Infinity [1] . . . . . . . . . . . . . . . . . . 25
3.12RIPComparison ............................ 26
3.13 OSPF Packet Format [8] . . . . . . . . . . . . . . . . . . . . . . . . 28
3.14 OSPF Neighbor Formation [8] . . . . . . . . . . . . . . . . . . . . . 30
3.15 OSPF Packet Types [8] . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.16LSATypes[8].............................. 32
3.17 OSPF Algorithm [8] . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.18 DR, BDR Selection [8] . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.19 OSFP Virtual Links [8] . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.20OSFPAreaTypes[8].......................... 36
4.1 NetworkDomain ............................ 38
4.2 NodeDomain .............................. 38
xii

4.3 ProcessDomain............................. 39
4.4 Flow Chart of Design . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.5 NetworkDesign............................. 40
4.6 ConvergenceTime ........................... 41
4.7 Routing Updates Traffic Sent . . . . . . . . . . . . . . . . . . . . . 42
4.8 Routing Updates Traffic Sent After Failure . . . . . . . . . . . . . . 43
4.9 Voice End-to-End Delay . . . . . . . . . . . . . . . . . . . . . . . . 44
4.10 Voice End-to-End Delay After Failure . . . . . . . . . . . . . . . . . 45
4.11 Voice Packet Delay Variation . . . . . . . . . . . . . . . . . . . . . 46
xiii

Chapter 1
Introduction
1

1
Introduction
1.1 Routing protocols
At present, communication networks are growing at a rapid speed. They fa-
cilitate users by providing different services like voice, video streaming and voice
applications. Internet is a major example of communication networks. Communi-
cation networks provide the basic infrastructure whereas routing protocols provide
the mechanism to exchange information among different nodes. Routing protocols
reside at network layer of Open System Interconnection (OSI) model. There are
two types of protocols working at network layer, i.e.,
•Routed Protocol
•Routing Protocol
Routed protocols carry the data like Internet Protocol (IP), Inter-network
Packet eXchange (IPX) etc., whereas routing protocols are used for the path selec-
tions like RIP, EIGRP and OSPF etc. Routing protocols can be further classified
into two major categories.
•Interior Gateway Routing Protocol (IGRP)
•Exterior Gateway Routing Protocol (EGRP)
IGRP is used for the route selection within the autonomous system. IGRP
can be classified into three major categories.
1. Distance Vector Routing (DVR) Protocol
2

2. Linked state routing protocol
3. Hybrid routing protocol (Advanced distance vector)
Bellman-Ford algorithm is used for path computation by DVR protocol. DVR
protocol periodically exchanges its complete routing table. DVR protocols broad-
cast its complete routing table to its neighboring nodes. Whereas link-state pro-
tocols use the triggered update instead of periodic updates. The DVR protocol
doesn’t use sophisticated process to compute the path. Linked state routing proto-
col uses the multicast to inform any change occurred in the topology. Linked state
routing protocols have much better self healing and path computational mecha-
nism.
Distance vector routers do not know about the complete network topology.
These routers only have information which is passed by the connected neighbor-
ing devices: Distance vector routing protocol only uses the hop-count as its metric.
Linked state and hybrid routing protocols also consider other parameters to com-
pute the path.
Figure 1.1: Distance vector protocol
In Figure 1.1, we have a topology in which three routers are connected with
each others. We have two routes for reaching A-C, one is directly from A-C and
the second route is A-B-C. If in this scenario RIP is used as a routing protocol
it will select A-B as the best path as according to Bellman-Ford algorithms only
considers the hop count, A-B has lesser hop-count as compared to path A-B-C.
Although A-B is only 56K link whereas A-B-C is 1 Gigabit links. If we implement
any linked state or hybrid routing protocol in this scenario it will always consider
path A-B-C, as this path have greater bandwidth as compared to path A-B. As
both the linked state and hybrid routing protocols mainly considers the bandwidth
3

for path computational and selection.
RIPv1 and RIPv2 and IGRP are the major example of distance vector routing
protocol. EGP and BGP are not pure distance vector routing protocols. A DVR
protocol calculates routes based solely on the link costs, whereas in BGP, the local
route preference value takes priority over the link cost.
Enhanced IGRP (EIGRP) is one of the major examples of hybrid routing
protocols. EIGRP is Cisco’s proprietary protocol. EIGRP provides much efficient
route computation as compared to DVR protocols.
1.2 Motivation
Packet Switched networks require much excellent, proactive and efficient rout-
ing protocol. To select the best routing protocol among RIP, EIGRP and OSPF
for the real time applications like streaming videos, IP telephony, IP unicast and
IP multicast traffic is the major motivation for this thesis. We have analyzed that
which routing protocol will perform better in different scenarios and environment.
1.3 Problem Statement
Routing protocols operate at layer 3 of OSI model. The basic purpose of net-
work layer is sorting and distribution of IP packets. There are several routing
protocols for dynamically selecting the best path. The main drawback of the
EIGRP routing protocol is that it only works when all devices are from Cisco.
Convergence time of EIGRP is faster and easy to configure as compared to other
routing protocols. In contrast, OSPF is a link-state interior gateway protocol.
OSPF is based on Dijkstra algorithm. OSPF routing protocol is difficult to con-
figure. OSPF requires high memory and high processor requirements.
Routing Information protocol (RIP) is not widely used routing protocol these
days, it’s only used in small network setup, and RIP uses a very simple routing
algorithms Bellman Ford algorithms. There are many drawbacks of RIP which
are discussed in chapter 3 and chapter 4. In this thesis, focus is on measurement,
architecture, performance and verification of the IGRP.
4

1.4 Thesis organization
We provide technical details about routing protocols in chapter 2. Chapter 3
presents a review of analyzed routing protocols. Chapter 4 presents an analytical
analysis of the selected protocols using OPNET simulator. We conclude the thesis
in chapter 5.
5

Chapter 2
Routing Protocols - Background
6

2
Routing Protocols - Background
2.1 Overview
In traditional packet switched networks, routing protocols usually transmits
packets routing information between interconnected nodes. In an IP network
routing decision takes place in hop by hop fashion. All the Routing protocols
must have these objectives:
1. To communicateobetween the different routers placed at different location.
2. To make correct and efficient routing decisions
3. To exchange information among neighbor’s routers.
4. To build error free routing tables
5. To learn existing routes
Routers are used to communicate among the different subnets and exchange
routing information among them. The main concept of ”routing protocols” is to
create the best possible path from one end to the other that is to find out the best
route from source to destination. These are built on basis of various properties of
the path. For IP based routing, Dynamic IGP protocols can be classified as;
•Linked State RoutingoProtocols
•Distance Vector Routing Protocols
•Hybrid.
7

2.2 Routing Protocols Attribute and Character-
istics
The main characteristics of routing protocols are given below:
2.2.1 Best Possible Routes
One of the characteristics of routing protocols is to find the best route possible
for communication. It might be one of the smallest or with least traffic depending
on the situation.
2.2.2 Faster Convergence
The communication time should be small in the inter-router communication
so information regarding routers can be easily acknowledged.
2.2.3 Security Options
The protocol makes sure that data is transmitted securely from source to given
destination.
2.2.4 Avoiding Loops
Loop free is used to find out the effective bandwidth of the network by making
sure that there are no loops in the network as loops would not only slow down but
would also make it tough to calculate effective bandwidth.
2.3 Important Concepts About Routing and Met-
rics
Understandings of these concepts are very essential before we start our study
about any IGP or EGP routing protocols.
8

2.3.1 Metrics
Metric can be simply define as decision making criteria of any routing protocol
on the basis of which it makes the routing decision is called as metric. All protocols
have different criteria for considering the best route. For example in case of RIP
hop-count is metric, EIGRP uses composite metric which is the combination of
different attributes like, load, reliability, MTU, bandwidth and delay, Whereas
metric of OSPF is cost based on bandwidth.
2.3.2 Need of a Metric
Metrics normally determine the best possible path in case when more than
one path is available for the same destination node. There are various ways to
compute best path and metrics for each routing protocol.
2.3.3 Parameters of a Metric
Every routing protocol uses Metrics to measure and then rank routes on the ba-
sis of; Best to worst of most preferred to least preferred. Various routing protocols
use various routing techniques and parameters to measure the different metrics.
Following metrics are usually used to determine best paths by any routing proto-
col.
•Delay: determines the amount of time needed to travel from one point to
other. It depends on various parameters, like; bandwidth utilization of link,
physical distance travelled, and port queues.
•Cost: can be based on any single metric or combination of metrics as it
can be determined by shortest distance or least traffic path. It is normally
administered by network administrators.
•Load: It is calculated by the traffic on the specific nodes. The routing
protocol normally uses loadoin the calculationoof a best route.
•Hop-count: It counts the number of routers for which a packet has to cross
to reach destination.
•Bandwidth: also plays an important role in the path identification which
normally results in choosing a high bandwidth link over a low bandwidth
link.
9

•Reliability: can be calculated by referencing to the earlier failure happenings
or previous error counts.
2.4 classification of Routing Protocols
Routing means best possible route. Routing protocols can be classified in to
two major categories that is State route vs. Dynamic route. State route are the
routes configured by network administrator. Static routes administrative distance
is always 1, means if there is static route in a network and some other dynamic
routing protocols are also used for the path selection that static route will always
be considered as a best route because it has lesser administrative distance as com-
pared with the any routing protocol. Dynamic routes are the routes selected using
routing protocol. Dynamic routes can be classified in to two major categories that
is interior gateway routing protocols and exterior gateway routing protocols. Inte-
rior gateway routing protocols are used for route selection within the autonomous
system the most popular IGP protocols are RIP,OSPF ,IS-IS and EIGRP. Each of
the routing protocol has their own metric and administrative distances. Exterior
gateway routing protocols are used for routing among different autonomous sys-
tem; the major example of exterior gateway routing protocol is Border gateway
routing protocols. Routing protocols can be classified in the following fashion.
Figure 2.1: Classification of Routing Protocols
10

2.5 Static and Dynamic Routing
In static routing, network administrator configures all the routes manually.
So, for every new addition network administrator has to add new routing entry
for the any addition. Whenever there is a new route or node added in the net-
work, manual entries have to be made and if node has to be deleted, it has to
be deleted manually as well. This is normally used for small networks. In static
routing, the network has more control over the network. Static routing is simple
to setup and less processor and memory intensive which is its one of the major
benefit but problem arises when there is change in the topology and reconfiguring
the route manually sometime become very problematic which is the drawback of
static routing. In static routing network administrator has much more control
and understanding of network because he is manually establishing the path and
knows which of his path is more reliable and what path to choose for specific type
of traffic. Whereas along with many advantages of dynamic routing protocol has
some problem as well, in dynamic routing, routing tables are formed dynamically
on the basis of routing protocol and its path calculation algorithms. The main
drawback of dynamic routing protocols is they are more processor and memory
intensive. Dynamic routing protocols have their own methods and technique of
establishing the best path. Normally linked state and hybrid of advanced distance
vector routing protocols maintains topology table in which all the possible path to
destinations are exist which is huge advantage which any network administrator
can utilize by the use of dynamic routing protocol.
2.6 Classfull Routing Protocols and Classless Rout-
ing Protocols
Based on the subnet mask, routing protocols are separated into two routing
protocols such as
•Classless Routing Protocols
•Classful Routing Protocols
2.6.1 Classfull Routing
There are three major classes of IP address
11

1. Class Aoranges from 1-126. In class A one octat is fixed for the network
portion where as three potions are fixed for the host portion. Subnet mask
of class A is 255.0.0.0.
2. Class Boranges from 128-191. In class B two octats are fixed for the network
portion and two of the octact are reserved for the host poritiono. Subnet
mask of class B is 255.255.0.0.
3. Class Coranges from 192-223. In Class C three octacts are reserved for the
network portions and only one ocatacts is reserved for the host portions.
•Routers use the same subnet mask to understand the network address whichois
directly associated to the interfaceoof the main network. When the router is
not directly associated to the interfaceoof the same mainonetwork, it applies
Classfull subnet mask to the route. Classfull routing protocols have so many
disadvantages and not used extensively in these days networks:
•This protocol doesn’t support Variable Length Subnet Masks (VLSM).
•They are unable to support discontinuous networks.
•These protocols cannot clinch routing updates.
•These protocol cannot be used in sub-netted networks
Below figure depicts the example of network in which class routing is deployed
with the same network mask in all the network locations.
Figure 2.2: Network Deployed with Classful Routing and Same Subnet Mask
2.6.2 Classless Routing
As there is limitation with the IPV4 addresses and there are so many reasons of
this shortage. Also commercial classes are not free of cost and anyone using these
addresses have to pay for it. So it will be very difficult for anyone to have public
12

addresses at each of the locations as IPV4 address allocation and distributions
is not good. Now days it has become a major quality of any routing protocol to
support variable length subnet masking. Below figure depicts the classless network
deployed by using variable length subnet masking at each of the location.
Figure 2.3: Network Deployed with Classless Routing Protocol
2.7 Linked-State Routing
The need to overcome the problem with the distance vector routing protocol led
to the development of linked-stateorouting protocol. Link-state routing protocol
have the following functionalities.
1. Respond quickly upon the change in the network, instead of periodic updates
triggered updates are used in case of link-state routing protocol.
2. Respond quickly on any network change.
3. Self healing on network route not working.
Link-state routing protocol only send update only when there is any change
occur in the network by this way not using the bandwidth unnecessarily. When
there is change in the status of any link it sends any advertisement known as link
state advertisement LSA.
By using LSA information is exchanged amongst all nodes. each LSA con-
tains the information of neighboring device. Any change in link is communicated
through LSA by flooding. All nodes can maintain a same database for all the
routes known as topology table. These databases provide in sequenceoinformation
of the link cost in the network. By this way routing table is formulated.” There
13

Figure 2.4: Link State Data Structure
is a routing table which includes information regarding link costs, their paths and
regarding all the neighboring nodes. Dijkstraoalgorithm is used forocalculating
the path andocost for each link. The linkocost is set by the networkooperator and
it isorepresented as theoweight or lengthoof that particular link.” Loadobalancing
performanceois achieved afteroassigning the linkocost. Therefore, link overcrowd-
ing of the network resources can be avoided. Therefore network operators can
change the routing by changing the link cost. Usually the costs of the link are left
with theodefault values and it is recommended to reverse the link’s volume and
then allocate the weight of a link on it.” Though link state request protocols have
better springiness, they are complex compared to the DV protocols.
Figure 2.5: Link State Routing [5]
There are two linked-state routing protocols that is OSPF and integrated IS-
IS. Link-state routing protocols use SPF (Shortest Path First Algorithm) knows
as Dijkstra’s Algorithm. Integrated IS-IS however uses SPF only.
14

2.7.1 Link-state protocol Characteristics
Link state protocol has the following characteristics.
•Each router possesses the identical database
•provides hierarchical structure
•Include and maintain several paths in the topology table for the destination
•Efficient and fast convergence without any loop
•Have much more precise metrics
2.7.2 Methods of Routing
These are the steps involved in the link-state routing.
•Every router acquires information of the directly connected neighboring net-
works and its directly connected links
•Every router stores information of link-state packet received from its neigh-
bor
•Every router establishes the minimum cost path for the network topology
•Every router should have a connection with its directly connected adjacent
networks and this is usually performed through ARP packet exchanges
•Every router must send a link-state information
2.7.3 Strength and Weaknesses of LSR
In Linked-state routing protocols, routersocalculate routes autonomouslyoand
are autonomous of theocalculation of intermediate routers. The strength of linked-
state routing protocols are:
•They act fast to any change in connectivity.
•The packet size is very small.
Major disadvantages of link-state routing protocols are:
•memory intensive
•Hard to configure and understand
•Uses sophisticated algorithms
15

•More processor intensive
•Usually Unsuccessful under agility for link changes
2.8 Distance-Vector Routing
The Major exampleoof Distance VectoroRouting Protocols is RoutingoInformation
Protocol (RIP), Distance Vector protocols as the name suggests considers the two
primary parameters to decide the best route; that are Distance and Vector. Dis-
tanceomeans ”How far” and Vector means which directions.RIP version 1 and
RIP Version 2 uses Bellman Ford algorithms. Administrative distance of RIP is
120.Administrative Distance is the degree of reliability of any routing protocols.
Smaller the administrative distance greater will be the reliability of routing pro-
tocols. Metric of Routing Information protocols is ”hop-count”. Metric is the
decision making criteria of any routing protocol. RIP is normally used for smaller
networks and its hop limit is 15. Distance vector routing protocols are also re-
ferred as routing by rumor. Distance vector routers have only information what
neighboring router had passed on. Distance vector routing protocols uses the pe-
riodic update instead of triggered update. There are so many problems associated
with the distance-vector routing protocol. One of the problems with the distance
vector routing protocol is that they are not aware of the full topology. Below
figure shows how the routes are learned by distance vector routing protocols and
it’s much easier to understand the problems associated with the distance vector
routing protocols like split horizon, route poisoning etc.
Figure 2.6: Distance Vector Routing [5]
16

2.8.1 Characteristics of Distance Vector Routing
The characteristics of Distance Vectorsrouting protocol are given below.
•Distance Vector routing protocol defines its routing table where all neigh-
boursoare directly connectedowith the table at a steady period
•New informationoshould put inoeach routing table instantlyowhen the routesobecome
unreachable
•Distance Vector routing protocols are easy and effective in smaller networks
and thus require little management
•Distance Vector routing is mainly based on hop counts vector
•The Distance Vector algorithm is iterative
2.8.2 Pros and Cons of DVR
The advantages of Distance Vector routing protocols are:
•Not using sophisticated algorithms, works very efficiently for smaller net-
works
•DVR are easy to implement
•Uses less memory and processor as compared to LSR
Distance Vector routing protocol experiences many problems like route poison-
ing, split horizon, counting to infinity and routing loops are the main drawbacks
of DVRs. DVR uses Bellman Ford algorithms which isn’t an efficient algorithms.
Some of the most popular problems with DVR are:
•Possibility of routing loops
•Uses periodic updates and takes time to converge.
•DVR protocols have hop-limit
•Bellman Ford is not using rich metrics like Linked-state protocols.
17

Chapter 3
Dynamic Routing Protocols
18

3
Dynamic Routing Protocols
3.1 Routing Information Protocol
The RIP (Routing Information Protocol) is dynamic and a one of the distance-
vector protocols which use the hop count as a metric. It was designed for smaller
IP networks. For routing updates Routing Information use UDP port 520. It’s
Calculates the best path by counting the hops. Like rest of the distance-vector
protocols its take some time to converge. RIP required less memory as compare
to the link state protocols. It is good for a small network limited to few subnets
and a small number of routers.
Figure 3.1: RIP-Ver1 Packet Format [1]
19

RIP go for 15 hops only. It cannot handle if number hops increases more than
that infect it will not work. Anything more than 15 hops away it will consider
unreachable if it found anything that is away more than 15 hops. This is way
RIP use for loop prevention. Routing Information Protocol is class-full routing
protocol. RIPv1 advertise all the networks that it knows about but without its
subnets. This option is available in updated RIPv2 and RIP next generation
(RIPng), defined in RFC 2080. RIP next generation (RIPng) is the extension of
RIPv2 for IPv6 support. Packet formats are shown in Figures.
Figure 3.2: RIP-Ver2 Packet Format [1]
Figure 3.3: RIPng Packet Format [1]
RIP is considered is a good solution for small consistent networks. For larger
network RIP will be considered more complicated and it will difficult for RIP to
20

run it in smooth fashion because it have to refresh it routing table after every 30
sec which is not good. RIP (Routing Information Protocol) exchanges complete
routing after specific. Routing loops is one of the major problems in distance
vector protocol. Below figure can best describe the possible loop occurrence when
we have 10.4.0.0 network goes down.
Figure 3.4: Routing Loops in RIP [1]
Above is the same condition in which packets travel across network without
reaching to the final destination.
21

3.1.1 Split Horizon
If router sends back information to router from its get already is one the reason
of loop occurrence. To avoid this Split Horizon is introduced. According to this
rule router will never send information back to the router interface from where it
learn already. Normally this happen when router A in a network sends some up-
date to the router B on network and router B didn’t get that information because
of some failure. Split Horizon Rule for 10.4.0.0:
R2 only advertises 10.3.0.0 and 10.4.0.0 to R1.
R2 only advertises 10.2.0.0 and 10.1.0.0 to R3.
R1 only advertises 10.1.0.0 to R2.
R3 only advertises 10.4.0.0 to R2.
Figure 3.5: Example of Split Horizon [1]
To remove above problem from network Split Horizon rule is introduced.
3.1.2 Route Poisoning
Router will consider the route fail if it is advertised with infinite metric (that
is metric: 16) instead of marking.
3.1.3 Poison Reverse
Poison reverse is a technique in which a router tells its neighbor routers that
one of the link or router is no longer exist of failed. To do this, the notifying
routers set the number of hops to the unconnected gateway to a number that in-
dicates ”infinite”. AS Routing Information Protocol allows only up to 15 hops so
22

Figure 3.6: Example of Route Poisoning [1]
by setting the hop count to 16 would mean that this particular link of route is ”in-
finite.” Network 10.4.0.0 goes down. R3 Poisons route with an infinite metric.R3
sends triggered Poison update to R2:
Figure 3.7: Poison Reverse [1]
23

3.1.4 Hold Down Timers
After a route poisoning, router stets a hold-down timer for that particular
route and if gets any update with some kind of better metric then the recorded
metric which was recorded earlier with in the period of hold down timer, In this
case the hold down timer will be removed and will be possible to send the data
across the network. . During this process hold-down timer, the ”possibly down”
will represent the ”downed” route in the RIP routing table. Its default value is
180 sec. In below situation Hold down timer will take place.
Figure 3.8: Hold Down Timers [1]
3.1.5 Triggered Updates
When in network any route failed and do not wait for coming next update,
instead send immediate update by listing poison route. R2 Removes Route:
Figure 3.9: Triggered Updates [1]
24

3.1.6 Counting to Infinity
It will go to maximum up to 15 hops. After 15 hops it will be considered
Unreachable.
Figure 3.10: Counting to Infinity [1]
Hop Count is 16, 10.4.0.0 is unreachable:
Figure 3.11: Example of Counting to Infinity [1]
3.1.7 Comparison
Factor used to determine that whether we go for RIP of other Distance vector
protocol include the following.
•What is the network size?
25

•Compatibility between routers models
•Administrative knowledge.
Figure 3.12: RIP Comparison
Above diagram can help to choose the Routing protocol for the parameter we
interested more.
3.2 Enhanced Interior Gateway Routing Proto-
col (EIGRP)
EIGRP is one of the major examples of Hybrid routing protocols sometime re-
ferred as advanced DVR. EIGRP is Cisco’s proprietary protocol. EIGRP provides
much efficient route computation as compared to Distance Vector routing proto-
cols. EIGRP is based on Diffusing update algorithm (DUAL). EIGRP maintains
three different tables. That is
1. Neighbor Table
2. Topology Table
3. Routing Table
3.2.1 Neighbor Tables
Neighbor table contains the information of neighboring devices.
3.2.2 Topology Table
EIGRP maintains the topology table; topology table contains all the informa-
tion of all the routes. For example if there are four five paths for the network
26

destinations, topology tables will have all the paths.
3.2.3 Routing Table
Routing table has the best path for the destination. For example if there are
four or five paths in the topology table only one route out of four, five routes will
be included in the routing table. EIGRP has many advantages over the link-state
routing protocols because of its architecture.
EIGRP is very efficient routing protocol, it has the best of both the distance
vector and link-state routing protocol. In EIGRP every router can perform route
summarization, EIGRP is much simpler to implement. EIGRP has also the con-
cept of Stub to restrict the unwanted routing updates. EIGRP uses very sophis-
ticated routing algorithm DUAL. Just like link-state routing protocols EIGRP
routing protocol is very memory and processor intensive.
3.2.4 EIGRP Features
1. EIGRP uses both the best of distance vector and liked-state routing proto-
cols.
2. EIGRP is Class-less protocol.
3. There isn’t any need of route redistribution with IGRP the predecessor of
EIGRP.
4. EIGRP metric and IGRP metrics are directly translate able.
5. Convergence is very fast.
6. Uses partial updates when needed.
7. Consumes less bandwidths (no broadcasts, no periodic updates, updates
contain only changes)
8. Supportssmultiple networkslayer protocols like IP,Apple Talk and IPX/SPX
etc. EIGRP Operation
3.3 Open Shortest Path First (OSPF)
There are two routing protocols that link-state OSPF and integrated IS-IS.
Link-state routing protocols use SPF (Shortest Path First Algorithm) knows as Di-
27

jkstra’s Algorithm. Integrated IS-IS however uses SPF only for finding all routers
in an area and runs Partial Route Calculation (PRC) algorithm for IP reach abil-
ity. On the other hand OSPF runs only Dijkstra’s Algorithm. OSPF have the
following detail in the packet.
Figure 3.13: OSPF Packet Format [8]
OSPF maintains three tables which help OSPF to work efficiently:
•Routing Table
•Topology Table
•Neighbor Table
3.3.1 Routing Table
Routing table is a table that keeps the track of all Subnets so that it can
transport packets to required destination. It uses the shortest path first (SPF)
algorithm to populate the routing table.
3.3.2 Topology Table
Topology Table is also known as Link-stat table. This table has the information
of every link in network. Keep the track of whole network. Below is the detail
view of Topology Table.
28

3.3.3 Neighbor Table:
Neighbor Table has all the neighbor detail which is running OSPF. It contains
all the required information about neighbor that is required for communication.
Updates in OSPF are triggered and incremental which means it will only send
update LSU packet when there is a change in a network. However, OSPF sends
complete routing table information to its like DV routing protocol after every
30min, also known as LS refresh.
3.3.4 OSPF Area Design and Principles
OSPF area design is hierarchical which means every area must connect to
Backbone Area 0. Routers that connect two areas are termed as ABR and routers
on which redistribution or other AS connect to be termed as ASBR. Routers in a
particular area maintain a same topology table that is, they have same link-state
database. Principle behind area design is to localize update within that area.
3.3.5 OSPF Neighbor Formation
OSPF neighbor formation is an eight step process.
1. OSPF router-id is determined. On an OSPF enabled router router-id is the
highest IP address on the physical interface, loopback IP-address if config-
ured beats the physical IP-address. Good design suggests that router-id
should be hardcoded under the routing process.
2. OSPF adds the interface that is directed by a network command, in its
link-state database.
3. DOWN state. In this state OSPF send hello packet on OSPF enabled in-
terfaces. Hello packet contains information like Neighbor, authentication,
Router-id, Area-id, Network mask, DR/BDR address, Hello and dead inter-
vals.
4. INIT state. In this state OSPF check for the fields in hello packet that
are Hello and Dead interval, Area-id, Authentication passwork and Network
masks. If any of them do not match the relation between routers keeps
bouncing between DOWN and INIT.
5. 2-WAY stat. In this OSPF checks the neighboring router-id and if they
are already neighbors the simply reset their dead intervals. If they are not
29

neighbor already then the process to next state.
6. EX START state. In this DBD packets or exchanged. DBD contains cliff
notes of all the networks that in the database. Master slave relation is formed
first and after that master router sends its DBD packet first.
7. LOADING state. In this state DBD are saved and acknowledged. Both
routers check for sequence number in DBD and check it against its database.
If a prefix is missing a LSR is sent to the other router, SLAVE sends LSR
first because it gets DBD first. LSU is sent back in response to LSR with
detailed information of the missing prefix.
8. FULL state. In this neighbors are synchronized and after that Dijkstra’s
algorithm runs.
Figure 3.14: OSPF Neighbor Formation [8]
30

3.3.6 OSPF Packet Types
•Hello Packet
•Database Description Packet (DBD)
•Link-State Request (LSR)
•Link-State Advertisement (LSA)
•Link-State Update (LSU)
•Link-State Acknowledgement (LSACK)
Figure 3.15: OSPF Packet Types [8]
3.3.6.1 Types of LSAs
•LSA TYPE 1: Within area local routes are represent by type 1. These are
just Hello packets that router exchange to each other after specific interval.
Light green arrows in figure represent LSA TYPE 1.
•LSA TYPE 2: These are the packets that DR send to other routers in the
Area. Light green arrows in figure represent LSA TYPE 2.
•LSA TYPE 3: ABR advertise these routes. These are inter-area routes. In
routing table they represent by O IA. Blue arrows in figure represent LSA
TYPE 3.
31

•LSA TYPE 4: ABR advertise Packets which is known is LSA type 4.It helps
other router to know router ASBR location. Green broken arrows in figure
represent LSA TYPE 4.
•LSA TYPE 5: Such types of routes are the external routes that are being
advertised by ASBR to an area. Purple arrows in figure represent LSA
TYPE 5.
•
•LSA TYPE 7: The Type 7 LSAs, NSSA, is when the Stub area is receiving
routes that are redistributed within its area. Red arrows figure represent
LSA TYPE 7.
Figure 3.16: LSA Types [8]
3.3.7 OSPF Cost
OSPF finds cost of a particular based upon its bandwidth. Formula for cost
calculation is 100/BW in Mbps, this means it has certain limitations e.g. Cost
for 10 Meg link will be 10 and for 100 Meg will be 1 and 1 Gig will also be one
1. Auto cost reference changes that phenomena and allows us to adjust the cost
formula.
32

Figure 3.17: OSPF Algorithm [8]
3.3.8 Understanding DR and BDR
OSPF elects DR and a BDR on a shared segment. This is done on purpose to
get rid of redundant network updates for a single update. DR is the router with the
highest priority. DR and BDR role is critical when we into NBMA where particular
router(hub) must be manually configured as DR. FULL relation-ship is formed
between DR/BDR and every other router in a shared segment.2-way relation-ship
is formed be other.224.0.0.5 and 224.0.0.6 are the two multicast address used to
send update. Non-DR send updates on 224.0.0.6 which DR listens to , DR then
sends that update on 224.0.0.5 which all routers listen.
3.3.9 OSPF Virtual Link
OSPF design requires all routers to be physically connected to Area 0. In this
situation the routes of that areas are not advertise to any other area. If we run
into a situation where our area does not connect to backbone we can use virtual
links or GRE tunnels. Virtual link increases the domain to Area 0. Virtual link
however, solve our problem but it is not a reliable solution so, re-designing should
be done.
3.3.10 OSPF Summarization
Summarization in OSPF is restricted to two either ABR OR ASBR. It allows
updates to be localized to a certain area. Summarization is done under the routing
process. We can manually change the cost of summary routes. AREA RANGE
XX command on ABR and SUMMARY-ADDRESS command on ASBR is used
33

Figure 3.18: DR, BDR Selection [8]
Figure 3.19: OSFP Virtual Links [8]
to summarize.
Redistribution on the other hand can be done on ASBR only. OSPF redis-
tributes on class full boundary so subnet keyword must be entered while redis-
tributing into OSPF. By default OSPF sets the cost of 20 and external type-2.
Route-map can be referenced to tag routes as well as their default parameters can
also be overridden.
34

3.3.11 OSPF Route Filtering
Filtering can be done using distribute-list on a router but this only removes
a route from its routing table, OSPF database still retains and forwards the info
to neighboring routers. Distribute-list if configured on ABR that connect area 0
to other areas removes a network not only from routing table but also link-state
database. Only scenario where distribute list can be configured in out bound
direction is ASBR for external redistributed routes, when it is done the routes are
also removed from link-state database. LSA Type-3 summary routes cannot be
filtered on an ABR. Database filtering can also is accomplished on a particular
interface.
3.3.12 OSPF Area Types
There are different areas in OSPF depending upon the design requirements of
a network.
•OSPF STUB AREA: LSA Type-4 and 5 routes will not be propagated into
a stub area. ABR Installs a default routes into a stud for the reachability of
external routes.
•OSPF TOTALLY STUB: LSA Type 3 in addition to type 4 and 5 is not
propagated into this kind of area. This area will only maintain intra area
routes and a default route injected from ABR.
•OSPF NSSA: LSA Type-4 and 5 are not allowed in this area but we can
redistribute into this area. External routes will be propagated as LSA Type-
7 (NS1 or NS2) and ABR converts those into LSA Type-5 (E1 OR E2) and
advertise to other areas. Default route in not advertised but, we can make
ABR advertise the default route.
•OSPF NSSA TOTALLY STUBBY AREA: LSA Type-3 in addition to LSA
Type-4 and 5 are not propagated into these areas. This area will only main-
tain intra area route plus LSA Type-7 route.
35

Figure 3.20: OSFP Area Types [8]
Chapter 4
Simulation Results
36

4
Simulation Results
4.1 Simulation Environment
In this thesis, network simulator, Optimized Network Engineering Tools (OP-
NET) modeler 14.0 has been used as a simulation environment. OPNET is a
simulator built on top of discrete event system (DES) and it simulates the system
behavior by modeling each event in the system and processes it through user de-
fined processes [23]. OPNET is very powerful software to simulate heterogeneous
network with various protocols.
4.1.1 OPNET Structure
OPNET is a high level user interface that is built as of C and C + + source
code with huge library of OPNET function [24].
4.1.1.1 Hierarchical Structure
OPNET model is divided into three domains [24]. These are:
1. Network domain: Physical connection, interconnection and configuration
can be included in the network model. It represents over all system such as
network, sub-network on the geographical map to be simulated.
2. Node Domain: Node domain is an internal infrastructure of the network
domain. Node can be routers, workstations, satellite and so on.
3. Process Domain: Process domainsare used tosspecify the attributesof the
processor and queuesmodel by using sourcescode C and C ++ which is inside
37

Figure 4.1: Network Domain
Figure 4.2: Node Domain
the node models.
38

Figure 4.3: Process Domain
4.1.2 How to Analyze and Design in OPNET
When implementing a real model of the system in the OPNET, some steps are
to be followed to design on simulator. Figure 5.1 shows a flow chart of the steps
[3].
Figure 4.4: Flow Chart of Design
4.1.3 OPNET Environment
We used OPNET Modeler version 14.0.A for network simulations. OPNET is a
comprehensive network simulation tool with a multitude of powerful functions. It
enables simulation of heterogeneous networks by employing a various protocols [2].
39

Figure 4.5: Network Design
We have simulated a real time network of Company XYZ. We have four offices
connected together in full mesh topology. We have followed the Cisco’s recom-
mended three layer hierarchical model. That is
•Core
•Distribution
•Access
Cisco 7609 router is used as Core router at each office. All these offices con-
nected using E1 connections. All the relevant configuration regarding the IP
configurations and also routing related configuration done and tested. At each
branch 6509 Core switch is deployed as Core Switch and 6506 used as aggregation
switches. We have used Cisco 2960 Switch as an Access Switch. All the access
switches configured with Hot Standby Routing Protocol to make sure the high
availability.
The first simulated network employs the RIP routing. The same model is then
used to simulate EIGRP and the OSPF routing protocol. The three scenarios
are: ”RIP no fail”, ”EIGRP no fail”, and ”OSPF no fail”. We added the fail-
ure/recovery setting (the link between Subnet1 and Subnet5 fails at 300 s and
recovers at 500 s) to each scenario and created three additional scenarios named:
RIP, EIGRP, and OSPF
40

4.2 Simulation Results
We simulated the network convergence activity and protocol traffic using six
simulation scenarios. The RIP, EIGRP, and OSPF protocol are chosen under
global statistics.
Network Convergence: Our study of network convergence scenario concludes
that EIGRP has the shortest convergence time whereas OSPF experience the
longest convergence period. RIP convergence time is better than OSPF. It shows
that EIGRP is much better due to its hybrid nature and DUAL algorithm. RIP is
better than OSPF because it has a simple design and limited functionality. OSPF
will take definitely need some time to calculate and convergence because it is
design for big networks and using SPF algorithm and have sophisticated selection
of DR and BDR in every LAN segment. The EIGRP and the OSPF protocol
experience the shortest and the longest network convergence times, respectively.
Figure 4.6: Convergence Time
EIGRP is very good in convergence because of its hybrid nature. OSPF take
more time than the rest of two because of its hierarchical structure and complex
criteria in selecting DR, BDR in every LAN segment. RIP need less time as
compare to OSPF because it is a simple protocol ever exist. It follows really simple
parameter and somehow its shows better result than OSPF. So above result can
help in selecting protocol as our design demand.
In case of no link failure in OSPF shows the most traffic sent. The reason of
this result is that due to packet header size and biggest conversion time OSPF
took more time to send traffic update as compared with EIGRP and RIP. EIGRP
41

is the fastest routing protocol used in this simulation to sent traffic update because
of its very fast convergence time; also RIP was the second to send traffic update
because of its better convergence time than OSPF.
Routing Updates Traffic sent (bits/sec)
EIGRP
OSPF
RIP
Figure 4.7: Routing Updates Traffic Sent
42

Same test was conducted with the node failure; with the node failure EIGRP
provided better results than that of OSPF.
Routing Updates Traffic sent (bits/sec)
EIGRP
OSPF
RIP
Figure 4.8: Routing Updates Traffic Sent After Failure
43

OSPF shown better results in case of delay variation. OSPF had better results
for the end -to-end delay and RIP shows the worst results for end-to-end delay.
EIGRP was slightly better than RIP for end-to-end delay. EIGRP delay was much
closed to the OSPF delay. EIGRP and OSPF better results than RIP.
seconds
EIGRP
OSPF
RIP
Figure 4.9: Voice End-to-End Delay
44

OSPF response after the link failure is less than the EIGRP.
seconds
EIGRP
RIP
OSPF
Figure 4.10: Voice End-to-End Delay After Failure
45

As for as delay variation results is concerned EIGRP and RIP shown almost
similar results unlike OSPF which shows higher delay variations in the beginning
but after the steady state achieved all these routing protocols shows similar sort
of results.
seconds
OSPF
EIGRP
RIP
Figure 4.11: Voice Packet Delay Variation
For the Voice packet traffic sent; voice traffic sent activity is not affected by
delay, jitter, and delay variation. All of these routing protocols shown similar
results.
46

Chapter 5
Conclusions
47

5
Conclusions
Interior dynamic routing protocols like OSPF and EIGRP are heavily used in
industry along with Routing information protocol (RIP). In this thesis we have
analyzed the behavior of all these routing protocols using different test scenarios
for the real time applications. Performance of all these protocols tested on the
basis of attributes associated with these protocols to figure out which of the routing
protocol best fit the industry requirements for the real time applications.
Enhanced Interior Gateway routing protocol (EIGRP) which is Cisco’s propri-
ety routing protocol has the best result in case of convergence duration and for
the same topology EIGRP was the first to converge. EIGRP has the best way out
in case of link or node failure, because of its topology database and concept of
successor and feasible successor. Therefore we can say that EIGRP is much reli-
able for the real time applications. For routing traffic Enhanced Interior Gateway
Routing Protocol was the first one to send traffic. Routing Information protocol
had the slightest traffic. Because RIP only sends the number of hops information.
Whereas, OSPF with the most traffic sent and was the last one to send routing
traffic. EIGRP has the mechanism to send traffic very fast after link failure as
compared with the OSPF. OSPF shown better results in case of delay variation.
OSPF had better results for the end -to-end delay and RIP shows the worst results
for end-to-end delay. EIGRP was slightly better than RIP for end-to-end delay.
One of the best things about OSPF is its area design and hierarchical structure,
by doing so we can manage the routing table size, less amount of route processing
and memory usage. On the contrary, areas increase the amount of configuration
and also very sophisticated configuration required to obtain the best out of it.
OSPF area design also help reducing the connectivity and increasing the traffic
concentration. OSPF is widely used interior gateway routing protocols because of
48

its design and the amount of liberty in design and enhances the network efficiency
by reducing the unwanted routing updates and by maintain the different tables.
49

References
50

References
[1] Rick Graziani and Allan Jonson, “Routing protocols and concepts: CCNA
exploration companion guide,” Pearson Education. London, 2008.
[2] Mohammad Nazrul Islam, Md. Ahsan Ullah Ashiqu, Simulation Based EIGRP
over OSPF Performance Analysis, Master Thesis in Electrical Engineering Em-
phasis on Telecommunications Thesis no: 4983 May 14, 2010
[3] Dong (Don) Xu, OSPF, EIGRP AND RIP PEFORMANCE ANALY-
SIS BASED ON OPNET, ENSC835: COMMUNICATION NETWORKS,
SPRING 2011
[4] Cisco Systems, I., Cisco Networking Academy Program CCNA 1 and 2 Com-
panion Guide Third Edition.
[5] Routing Overview available at: http://networking.ringofsaturn.com/IP/Routing.php
[6] Moy, John. OSPF Anatomy of an internet routing protocol. May 200.
[7] Tanenbaum, Andrew s. Computer Networks. s.l.: Pearson Education, 2003.
[8] Berkowitz,Howard.http://www.certificationzone.com/cisco/studyguides/component.html?module=studyguides.[Online]
OSPF Part 2: Using OSPF in Hierarchical Systems.
[9] ] IKram Ud Din, Saeed Mahfooz and Muhammad Adnan ,Analysis of the
Routing Protocols in Real Time Transmission, Global Journal of Computer
Science and Technology, Vol. 10 Issue 5 Ver. 1.0 July 2010 p.18-22
[10] ] Network Working Group at: http://www.ietf.org/rfc/rfc2328.txt
[11] Internetworking Technologies Handbook Fourth Edition
[12] Mohammed A. Aabed RoutingInOPNET(30/11/2008)
51