Behavior-based Attack Detection and Reporting in Wireless Sensor Networks

Providing security for wireless sensor networks is challenging due to the resource limitations of sensor nodes. Leveraging the broadcast nature of wireless communication, we propose a behavior-based approach for attack detection and reporting in WSNs. In our approach, some ADS capable nodes are chosen for judging the behavior of their neighbors and conveying alarms to an administration sink when attacks are detected. We prove that the optimal selection of ADS capable nodes for minimum resource consumption is a NP-hard problem and present two heuristic algorithms for it. We also prove the effectiveness of our approach through extensive simulation experiments.