Misbehaving Node Detection in Mobile Ad Hoc Networks using Multi Hop Acknowledgement Scheme

Abstract

Problem statement: For any node in a Mobile Ad hoc Network conservation of battery power and bandwidth are the priority. Hence, they try to reduce the overload they would otherwise incur when they forward packets. This selfish behavior of a node affects the throughput of the network. The nodes may also choose a back off value of shorter duration. These problems are handled effectively by the methodology proposed in this study. A conscious effort has been made keeping the constraints of the MANETs in mind. Approach: Misbehavior is best identified at the lower levels as the upper levels of the OSI standard primarily deals with the data the packets carry and less about how, so the network and the MAC layers is where the primary focus lies. Thus, keeping these points in mind, we propose algorithms that work along with the 802.11 MAC protocol to monitor the behavior of neighboring nodes by listening to the channel, specifically monitoring parameters like back off values sent by the nodes. A counter is maintained which is incremented every time node misconduct is detected, subsequently after a particular value is cross the node is labeled as misbehaving and the information is broadcast over the network. Results: Performance parameters like throughput, packet delivery ratio were monitored with traffic of the magnitude 10 to 60 nodes. Also the performance of the network based on the percentage of selfish nodes present in the network was monitored and a graph was generated based on the statistics. Conclusion: An algorithmic approach for misbehaving node detection and isolation in ad hoc networks by modifying the protocol being used in the lower layers which consequently improves performance of the network had been proposed. Simulation results show considerable performance increase upon implementing the proposed algorithm. Further research can confirm the practicality of the proposed idea.