Journals Proceedings

Abstract

An important problem in wireless network security is node authentication. The existing schemes for node authentication are based on IP and MAC address which can be easily spoofed by the adversaries. Wireless spoofing attacks are easy to establish and significantly affects the functionality of the network. In the proposed approach, log files are used for authenticating each node which is hard to falsify and difficult to modify or delete the contents. The proposed system determines the spoofing attacks and localize the adversaries using spatial correlation of received signal strength (RSS). In this paper, the problem of identifying the number of attackers is determined by Partitioning Around Mediods (PAM), a cluster-based analysis method and Euclidean distance between successive packets. The simulation results show that the proposed approach over performs all other existing approaches with high detection ratio.