AN EXTENDED AD HOC ON-DEMAND DISTANCE VECTOR ROUTING PROTOCOL FOR PERFORMANCE IMPROVEMENT IN MOBILE AD HOC NETWORKS

Abstract

Mobile Ad hoc Networks have unique characteristics such as dynamic topology, ease of deployment, self-configuration, they require no prior laid down infrastructure, self-organisation, and lack of centralized control. However, devices participating in MANETs face challenges including routing overheads, link breakages, security threats, nodes limited battery power, packet loss, limited bandwidth, scalability problems, transmission range constraints and memory constraints. Despite the development of various AODV variants with an aim of reducing the routing control overheads in MANET, flooding remains a challenge. The routing overheads cause network performance degradation due to potential of collisions and congestion inMANETs. The main goal of the research was to extend Ad hoc On-Demand Distance Vector (AODV) routing protocol to cause performance improvement in MANETs.The study sought to analyze the characteristics of AODV with an aim of extending it for performance improvement in MANET, and design, develop and evaluate the extended AODV protocol through simulation experiments. The extension of AODV protocol was done by integrating passive clustering with energy awareness within its route discovery logic, thus reducing the routing overheads through minimizing the number of nodes rebroadcasting the route requests and creating a stable network topology. The study was guided by network and graph theories. The evaluation of the extended AODV protocol (X-AODV) against benchmark variants was done using routing overhead, throughput, packet loss, end-to-end delay and packet delivery ratio performance metrics. The study employed pragmatism philosophy and design science methodology for X-AODV development. Objective one of study was achieved through a systematic review and an expert opinion survey, while objective two through protocol architecture design and algorithm design, objective three coding and simulation experiments while objective four was achieved through simulation experiments. The X-AODV protocol was simulated in Network Simulator 2, using fifteen MANET scenarios with a network with nodes ranging from 10 to 150. The simulation data was analyzed, evaluated and output graphs generated using Gnuplot software. The performance of X-AODV, was compared with conventional AODV, Energy Aware AODV and AOMDV as benchmark variants. The simulation results indicated that X-AODV protocol registered a steady rise in throughput ranging from 3.05 packets per second at a node density of 10 nodes to 89.40 packets per second at a node density of 150 nodes. Benchmark variants throughput gradually declined to 14.94 packets per second for AOMDV, 25.51 packets per second for AODV and 24.66 packets per second for EA-AODV at a node density of 150. Further, X-AODV end￾to-end delay was below one second, ranging from 0.241 seconds to 0.895 seconds,while the variants had an end-to-end delay of up to 4.239 seconds. The extended AODV protocol had its count of routing overhead drastically reduced, ranging from 350 to 17593 packets, while the benchmark variants had a count of up to 1,341,821 packets at node density of 150. X-AODV had packet losses of up to 79%, while the benchmark variants had losses of up to 96%, while its packet delivery ratio was stable between 21% and 26%, but the benchmark variants had the packet delivery ratio drop drastically to a low of 3.29% for AOMDV, 5.61% for AODV and 6.3% for EA-AODV at a node density of 150. Therefore, X-AODV performance was superior to the benchmark variants, thus it was scalable, and it had a high routing efficiency with faster delivery times and superior real time performance