Mobile Agents Based Congestion Aware Routing in Mobile Ad Hoc Networks

This paper proposes a mobile agent based congestion aware routing (MACAR) protocol, for routing in mobile ad hoc networks (MANETs). Here mobile agents collect network connectivity information based on congestion state and disperse this information across the network. This allows the nodes in the network to make "congestion aware decisions" during route discovery and route maintenance phases. The congestion state is measured based on the interface queue size and MAC drops at each node. In traditional on-demand routing protocols, "shortest path" is used as the metric in choosing and maintaining routes. MACAR chooses routes based on congestion as the primary metric. This results in a spatial distribution of traffic in the network and reduces congestion in the network. MACAR uses ad hoc on-demand distance vector (AODV) routing as the underlying protocol (C.E. Perkins). MACAR and AODV protocols are compared through extensive simulations. Network throughput, end-to-end delay and routing overhead are the performance parameters considered. Results show that the robust routes provided by MACAR when compared to traditional on-demand routing protocols results in substantial reduction in end-to-end delay and improvement in network throughput.