Coping with Traitor Attacks in Reputation Models for Wireless Sensor Networks

Most of the current reputation models for wireless sensor networks (WSNs) do not consider seriously the vulnerabilities of reputation models themselves, which makes them easily manipulated by traitor attacks. The traitors can launch attacks against reputation models on both service level and recommendation level. Until now, recommendation traitor attack has received little attention, although it causes great harm to reputation models. In this paper, we build an economic Cobweb model for traitor attacks on both of the two levels to exactly formalize traitors´ behaviors. Then we propose an attack-resilient scheme inspired by Macroeconomic Control. Through reputation fluctuation analysis, the traitors can be discriminated from benign sensors and then punished. In the simulations we verify that although both traitor attacks cause great harm to reputation models, our scheme can discriminate traitor sensors from benign sensors and resist traitor attacks effectively with lightweight energy consumption.