Module

8
Routing

Version 2 ECE, IIT Kharagpur

Communication networks generally have multipath between a source-destination

pair to transport the message between them. Devices are designed and
deployed to carry out the function of allocating a particular path, out of the
possible alternative paths, to a call based on various considerations. The path
can be circuit or packet switched. In circuit switching the path is chosen by the
communication network like the telephone system and switched (dedicated to the
call end to end before the actual message communication starts). In packet
switched networks, a path is shared by many users. Packets of information from
different users can travel through the same link. There are two modes of packet
switching. In the first mode every packet carries the source and destination
addresses and forwarded to an out going link by the packet switching node
following some algorithm. The process continues till the packet reaches the
destination or gets aborted due to various problems. In the second mode, a path
is first decided and then all the packets for a particular call follow the same path
one after another as and when the forward links are available. The packets may
have to wait at nodes. Selection of path and switching of the packets on the path
are called routing. Routing is corresponding to switching in circuit switched
telephone networks. Unlike in packet routing, one the call is set-up, there is no
further switching involved at the nodes during the progress of the call in circuit
switching. Thus routing is a some what more complex process than switching.
Obviously routing is vary crucial in computer networks.
Broadly routing can be static or dynamic. The routes are pre-decided in static
routing. In dynamic routing, the routes are influenced by the current traffic
conditions. To avoid congestion and to optimize resources utilization, many
routing techniques have been developed. Some of them are briefly discussed in
the following lessons.
A simple example of routing alternatives in circuit switching is shown in Figure
8.1.1. Similarly Figure 8.1.2 shows the example of routing in packet switching
networks. The least cost path is AEDC whereas paths with minimum number of
hops are ABC and AEC between the source A and destination C.
Requirements of routing include optimality, efficiency, stability, robustness,
simplicity, accuracy and fairness.

Version 2 ECE, IIT Kharagpur

Figure 8.1.1 Example of routing in circuit switched network

Paths from A to C
ABC
AEDC
AEC
ABEC
ABEDC

Number of Hops
1
2
1
2
3

Total cost
6
3
5
9
7

Figure 8.1.2 Routing example in packet switched network

Version 2 ECE, IIT Kharagpur

Lesson
26
Routing I

Version 2 ECE, IIT Kharagpur

Objective
To learn some routing techniques and some routing algorithms.

8.1.1 Introduction
The main function of the network layer is routing packets from the source
machine to the destination machine. In most subnets, packets will require
multiple hops to make the journey. The routing algorithm is that part of the
network layer software which is responsible for deciding on which output line the
incoming packet should be transmitted. This decision has to be made for each
packet in case of Connectionless Packet Switching and once per session for
Connection Oriented Service. Regardless of whether routes are chosen
independently for each packet or only when new connections are established,
there are certain properties which are desirable in a routing algorithm :
correctness, simplicity, robustness, stability, fairness and optimality. Robustness
implies that the routing algorithm should be able to cope with the changes in the
topology and traffic without requiring all jobs in all hosts to be aborted and the
network be rebooted every time any router crashes. Stability implies that a
change should not be back and forth in a short period of time. Other parameters
are easy to appreciate.

8.1.2 Routing Techniques

Routing requires a router to have a routing table. Several techniques are
employed to make the size of the routing table manageable and handle issues
such as security. Some of these techniques are discussed here.
NEXT-HOP ROUTING
In this technique, the routing table only holds the information that leads to the
next hop instead of holding the complete information.
NETWORK-SPECIFIC ROUTING
In this method instead of having an entry for every host connected to the same
physical network, we have only one entry to define the address of the network
itself. We treat all hosts connected to the same network as one single entity.
HOST-SPECIFIC ROUTING
In this technique the destination address of the host is given in the routing table.
This concept is opposite to the network-specific routing technique. This leads to
loss of efficiency. This method is used for specific purposes such as checking the
route or providing security measures.

Version 2 ECE, IIT Kharagpur

8.1.3 Routing Algorithms

The routing algorithm is that part of the network layer software responsible for
deciding on which output line an incoming packet should be transmitted. If the
subnet uses datagrams internally, this decision must be made a new for every
arriving data packet since the best route may have changed since last time. If the
subnet uses virtual circuits internally, routing decisions are made only when a
new virtual circuit is being set up. Thereafter, data packets just follow the
previously established route. The latter case is sometimes called session routing,
because a route remains in force for an entire user session (e.g., a login session
at a terminal or a file transfer)
Routing algorithms may be classified into two major classes : non-adaptive and
adaptive, depending on how the routes are calculated.
Non-adaptive algorithms, also known as Static algorithms do not base their
routing decisions on the estimates of current traffic and topology. Instead the
route is pre-computed and fed into the routers offline.
Adaptive algorithms on the other hand change their routing decisions to reflect
the changes in the topology and usually in traffic as well. The various adaptive
algorithms differ in where they get their information (from adjacent routers or from
all routers), when they change the routes (when the load changes or when the
topology changes), and what metric is used for optimization (distance, number of
hops, residual bandwidth). These algorithms are also known as Dynamic
algorithms.
The routing may be performed for the Uncials, one source to one destination or
Multicast, one source to a group of destinations scenario. In uncial routing, when
a router receives a packet it forwards the packet only through one of its ports (the
one belonging to the optimum path). The network may be so large that a single
routing protocol may not be able to handle the task of updating the routing table
of all the routers. So we divide the internet into Autonomous systems (AS).
Routing inside an autonomous system is interior routing. Routing between
autonomous systems is exterior routing. Each autonomous system can choose
an interior routing protocol to handle routing inside the autonomous system.
Several routing protocols are in use. Multicast routing requires group
management. Mechanisms are required to create and destroy groups. Processes
need to be developed where they can leave or join groups. The routers are
required to know which hosts belong to which groups. Same of the popularly
used routing protocols are discussed in the following lesson. Only nicest routing
is covered in this course.

Version 2 ECE, IIT Kharagpur

Questions
1. What is routing?
2. What are the requirements of a routing technique?
3. Differentiate between the minimum hop routing and minimum cost
routing.
4. Explain interior and exterior routing.

Version 2 ECE, IIT Kharagpur