Robust and Secure Time-Synchronization Against Sybil Attacks for Sensor Networks

Abstract

Time synchronization is crucial for cyber-physical systems (CPSs), e.g., wireless sensor and actuator networks. Cyber physical security is an important yet challenging problem. In particular, attacks to the time synchronization service may incur data distortion or even malfunction of the whole system. Sybil attack is one of the most common attack types for sensor networks, where a node illegitimately claims multiple identities. Existing secure time-synchronization protocols; however, cannot well address Sybil attacks in the time synchronization process. We propose a robust and secure time-synchronization protocol (RTSP) to defend against Sybil attacks. Different from previous secure timesync protocols, RTSP employs a novel graph theoretical approach, which is able to perform anomaly detection at the message level instead of the node level. This fine-grained detection ability enables RTSP to become robust against Sybil attacks, as well as node compromise and message manipulation attacks. Extensive experimental results show the effectiveness of our proposed protocol.