0
5.9kviews
What are the demerits of Distance Vector Algorithm? How is it overcome?

Mumbai university > Electronics and telecommunication Engineering > Sem 6 > Computer Communication and Telecom Network

Marks: 10

Years: May 2016

0
164views

• Distance vector routing is so named because it involves two factors: the distance and the vector i.e. the direction that it takes to get there.

• Routing information is only exchanged between directly connected neighbours.

• The protocol is simple, requires little management and is efficient for small networks. But they have poor convergence property.

• We consider two assumptions:

Each node knows the cost of the link to each of its directly connected neighbours.

A link that is down is assigned an infinite cost.


• The problem with distance vector routing is its slowness in converging to the correct answer. This is due to a problem called count to infinity problem.

• Another problem is that this algorithm does not take the line bandwidth into consideration when choosing root.

• The distance vector routing works properly theoretically but practically it has a serious problem called count to infinity problem. Although we get a correct answer, we get it slowly.

• Consider a topology with four nodes where each of them is connected to another and the link cost is 1.

• If the link between the 3rd and the 4th node breaks, then the algorithm tries to count till infinity. When the link breaks, node 3 counts that the shortest path to 4 is through 2 and it sends the data to 2.

• When node 2 receives data, it counts the shortest path to node 4 and finds it through node 1. Hence it sends data to both 3 and 1.

• Thus all the nodes believe that the neighbouring node has the shortest possible path and they update their routing table accordingly. The counting would not stop until the router runs out of the memory.

• The problem of count to infinity can be avoided by following a few mechanism namely:

o Defining maximum count: The routing information protocol defines a maximum count of 15. Hence the count to infinity would stop at 16th iteration. It also means that it could not support networks that are more than 5 hops away.

o Split horizon: It defines that the route does not send routing information back along the path on which the packet travelled.

o Poisoned reverse: It is an improvement over split horizon wherein the router sends data but puts the cost of the link to infinity.

o Hold down timers: It is the timer which is started when:

a)  Timer for particular route expires.

b)  A reachable route changes to unreachable route in an incoming response. When the timer reaches zero value,  the router is removed since the path is considered as dead or removed.