Controlling Congestion in Far-End Nodes

Estimating and attempting to overcome congestion in high radix topologies that use global adaptive routing decisions will have a great impact on the overall performance a heterogeneous network. Earlier works have been done focusing on near end congestion or downstream congestion, thus not evaluating far end congestion possibilities. In this project routing algorithms with near-end and far-end congestion metrics have been combined with a history window technique that helps to overcome far-end congestion. In addition to this, transient congestion would also occur on varying traffic patterns, which will be handled by a method of averaging local queue occupancies with an addition of offset. This combination of methods will get us closer to an ideal adaptive routing algorithm. Routing is the key functionality for directing communication over large networks. The primary task of any routing protocol is to discover and maintain routes to needed network destinations. The routing protocols for ad hoc networks can be divided into two groups, proactive and reactive. Proactive routing refers to the condition that whenever a node has some data for a particular destination it can be transmitted immediately. On the other hand reactive routing protocol determines the routes as and when it is required by a node in the network. Surprisingly, little focus has been directed to the enhancement or extending this approach.