Attack-resilient time synchronization for wireless sensor networks

The existing time synchronization schemes in sensor networks were not designed with security in mind, thus leaving them vulnerable to security attacks. In this paper, we first identify various attacks that are effective to several representative time synchronization schemes, and then focus on a specific type of attack called delay attack, which cannot be addressed by cryptographic techniques. Next we propose two approaches to detect and accommodate the delay attack. Our first approach uses the generalized extreme studentized deviate (GESD) algorithm to detect multiple outliers introduced by the compromised nodes; our second approach uses a threshold derived using a time transformation technique to filter out the outliers. Finally we show the effectiveness of these two schemes through extensive simulations.