Robust Sybil Detection for MANETs

In this research, we propose a robust Sybil attack detection framework for MANETs based on cooperative monitoring of network activities. We do not require designated and honest monitors to perform the Sybil attack detection. Each mobile node in the network observes packets passing through it and periodically exchanges its observations in order to determine the presence of an attack. Malicious nodes fabricating false observations will be detected and rendered ineffective. Our framework requires no centralized authority and, thus, is scalable in expanding network size. Privacy of each mobile node is also a consideration of our framework. Our preliminary experimental results yield above 80% accuracy (true positives) and about 10% error rate (false positives).