DESIGN AND ANALYSIS OF A MOBILE AGENT BASED ROUTING FOR MANETS

Abstract

Mobile Adhoc Network (MANET) is a collection of wireless mobile nodes that communicate over radio links without any existing infrastructure and central administration. They have dynamic topology and are bandwidth constrained. The biggest challenge in them is to find a path between the end points, which is aggravated through the node mobility. There are many existing proactive and reactive routing approaches for ad-hoc networks. Adhoc on Demand Vector (AODV) is one such reactive protocol, which is most efficient for routing in MANETs. Mobile Agents based approach can also be adopted for routing in MANETs. Mobile Agent is a software program that can suspend its execution on a host computer, transfer itself to another agent-enabled host on the network, and resume its execution on the new host. In this report, we have discussed Ant Routing Algorithm .(ARA) and proposed some modifications to ARA. We have simulated it on a discrete event network simulator to reduce the overhead and to discover the shortest route in MANETs. ARA is a mobile agent based protocol. We have simulated ARA in adhoc environment (Random way point model) having 15 nodes, 6Mbps data rate, 3 Constant Bit Rate (CBR) connections and 0.1 ms to 0.5ms link delay and found ant behavior is flexible enough to suit the dynamic nature of MANETs. In ARA, the ants are used for updating the routing tables of network nodes. AODV is also simulated in the same environment. The performance of ARA and AODV protocol are analyzed on different network parameters like delivery rate and overhead. It has been observed that modified ARA considerably reduces the overhead in MANETs. We have carried simulation on Linux-8.2 with gcc compiler and a discrete event network simulator OMNeT++ ver. 2.2 and 2.3. Intel Pentium IV 2.4 GHz. with 256 MB DDR and 133 MHz bus is used as system configuration. We have used Network Descriptor,(NED). Language, tcl/tk, C++ for Linux as languages and graphs are generated in Plove and MS Excel.