Scalable and Distributed Detection Anaysis on Wormhole Links in Wireless Sensor Networks for Networked Systems

Recent surge in cyber physical systems (CPSs) which integrate cyber and physical worlds, increases security risks at both the surfaces. The wireless sensor communication network is one of the major components of the critical infrastructure in all networked systems including networked control systems and CPSs. As these systems support range of time and data sensitive applications, varying from health care to emergency services, to transport, and to environmental monitoring services, core sensor network component faces potential threats due to varying attacks, including wormhole attack that can cause damage to the control system in terms of data leakage, data dropping, and delayed delivery. It is also difficult to detect, as it requires neither nodes in the network to be compromised nor adversaries to acquire valid network identity to instigate such attacks. Many existing solutions in the literature to detect wormholes are based on node's static location information, synchronization of clocks, use of additional hardware, such as antennas and GPS, and neighborhood and traffic information, which may lead to large energy consumption. In this paper, a scalable and distributed scheme which uses sequential probability ratio test is proposed, to avoid single point failures and to handle high mobility, with no additional resource requirements. It is observed that wormholes are detected using few packets and detection is faster with increasing mobility. The system is highly customizable as system parameters can be chosen to balance the speed and accuracy of detection. System overheads in terms of communication, computation, and storage aspects are analyzed and presented.