SybilVote: Formulas to Quantify the Success Probability of Sybil Attack in Online Social Network Voting

Abstract

This letter proposes to derive the success probability of Sybil attack in online social networks with the multiple-choice majority voting. The resultant SybilVote formulas produce outputs that are consistent with the Monte Carlo simulation and more accurate than the existing formula based on the multinomial distribution tail estimate. The computational complexity of SybilVote exact formulas is $O({{{({n + S})}^{k}}})$ , where $n$ , $k$ , and $S$ are the number of real users, choices, and Sybil users, respectively. The accurate approximation formula of SybilVote is also presented with $O(k)$ complexity by using a normal distribution approximation available when $n$ and $k$ are large in a large-population user voting condition. Finally, the effects of parameters on the success probability of Sybil attack have been investigated to highlight usefulness of the formulas.