# What is an Adaptive Routing Algorithm in Computer Network?

This Algorithm changes its routing persistence to pursue changes within the topology and traffic also. Routing decisions are often changed when network topology and load change. Adaptive routing aims to assist avoid packet delivery failure, improve network performance and relieve network congestion.

Adaptive routing can generate nodes to develop into overloaded because of the complex processing decisions they create. Because routers send data about the network topology, adaptive routing is often less ensure than non-adaptive routing processes and needed more bandwidth.

It is equivalent to GPS, which uses data about road conditions to redirect drivers, adaptive routing uses data about network congestion and node accessibility to direct packets. When a packet appears at a node, the node uses data sent among network routers to compute which direction is most accessible. If the default direction is congested, the packet is transmitted along with multiple directions, and the data is transmitted among network routers.

There are various types of Adaptive routing algorithm, which are as follows −

## Isolated

Each router creates a decision only for the local data it has on hand. Routers don’t even exchange information with their neighbours. The drawback of this is that it can share the packets through a congested path resulting in a delay. Some of the examples of this type of algorithms are

## Hot Potato Routing

A routing algorithm is the one in which routers of a network have no buffer to save packets before transferring to their last predetermined destination. In a normal routing situation, when the multiple packets go for a single outing channel, the packets that are not buffered are dropped to prevent congestion. The packet is bounced everywhere like a "hot potato" may be moved far away from their destinations. This algorithm doesn’t drop the packets but keeps on moving.

## Backward Learning

In this algorithm, decisions are made by optimally reaching. For example, assume a packet from node A can optimally get node C through node B. Each packet is attached with the originating source node identified and a hop counter that is incremented each time on every Hop.

When the node gets a packet in a specific border, it notes down the several hops it has taken to get it from the source node. If the previous value of Hop count is not better, then it is updated with new information. Else, don’t update. But if an optimal best route goes down, it can’t recall previous information which has been updated.

## Centralized Routing

The network’s central node takes complete data about the network topology about traffic and avoids several congestion communication nodes. This can transmit this data to the specific routers. The advantages of this are that only one node is required to keep updated information. But the disadvantage is that as the central goes down, the entire network falls.

## Distributed Routing

This node receives data from its neighbouring nodes and then decides which technique to transmit the packet. The drawback is that if it receives data between the intervals and sends the packet, something changes, and then the packet is delayed.

Updated on: 05-May-2021

1K+ Views