Sybil attack methods and detection technique on Kad network

Peer-to-peer(P2P) is a free and open distributed network. Every node could join and leave freely. Lack of verification of joining nodes will cause Sybil attack. Kad is the most popular P2P file-share system. By analyzing Kad protocol and software implementation, we find that several main processes have Sybil vulnerabilities, such as bootstrap, routing update and resource search. Based on these studies, this paper design Kad Sybil attack detection algorithm KSybil. It uses AGNES clustering algorithm to detect suspected Sybil groups on specific keywords. The experiments show that KSybil could detect Sybil attacks in Kad network effectively.