Design of an Agent Based Analysis Approach for Secure Communication Against the Existing of Selfish Node Attack

Abstract

The internal participating nodes in mobile network can be affects fully or partially by selfish nature to improve its own communication rate. This nature can degrade the forwarding rate and affects the overall performance of the network. The detection of such selfish nodes is required to improve the communication throughput and network performance. In this paper, an Agent based analysis approach is defined to detect the selfish node in the network. The fuzzy based zonal analysis method is suggested to perform the zonal evaluation on the mobile network. The distributed Agents performed the analysis on the fuzzy based communication parameters. The parameter specific evaluation method is able to separate the good and selfish nodes. Once the selfish nodes are identified, the preventive route formation is done for safe communication. In this paper, the Agent adaptive model is presented for detection and prevention of selfish nodes. The proposed Agent adaptive model is simulated in NS2 environment. The comparative results are generated against token based, agent based and watchdog methods. The analysis results show that the proposed protocol has achieve the 17% improvement in packet communication, 37% reduction in communication loss and 53ms reduction in packet delay.