CMNTS: Catching malicious nodes with trust support in wireless sensor networks

Abstract

Security in wireless sensor networks is critical due to its way of open communication. In this paper we have considered suite of attacks - packet modification, packet dropping, sybil attack, packet misrouting, and bad mouthing attack, and provided a solution to detect attacks. In literature, many schemes have been proposed to mitigate such attacks but very few detect the malicious nodes effectively and also no single solution detects all attacks. In the proposed approach, each node chooses the parent node for forwarding the packet towards sink. Each node adds its identity and trust on parent as a routing path marker and encrypts only the bytes added by node in packet before forwarding to parent. Sink can identify the malicious node based on trust value and node identities marked in packet. Child node observes the parent and decides the trust on parent based on successful and unsuccessful transactions. Data transmission is divided into multiple rounds of equal time duration. Each node chooses the parent node at the beginning of a round based on its own observation on parent. Simulated the algorithm in NS-3 and performance analysis is discussed by comparing the results with other two recently proposed approaches. With the combination of trust factor and fixed path routing to detect malicious activity, simulation results show that proposed method detect malicious nodes efficiently and early, and also with low percentage of false detection.