Currently, most of the existing social networks on Internet are distributed, decentralized systems, and they are particularly vulnerable to Sybil attack in which a single malicious user introduces multiple bogus identities and pretends to be multiple and real users in the network. With these controlled identities, the malicious user can create a Byzantine failure in collaborative tasks by ‘out vote’ the real identities. This paper conducts a survey on the network security of social networks to provide the overview of the current online security of the social networks and the corresponding defend methods. Based on the survey, this paper proposes Sybil-Resist, a
-based Sybil attack defense protocol devoting to identifying the Sybil nodes and the Sybil region efficiently. The simulation results obtained by a more realistic simulation topology show that the proposed scheme outperforms existing solutions in terms of detection accuracy and running time.