An efficient privacy-preserving P2P protocol for computing maximum value in the presence of active adversaries

Abstract

Joint and secure computation of the private data inputs of a group of users is an interesting problem in current P2P applications. The original problem of this type is the Millionaires’ Problem, in which two millionaires wish to know who is richer without disclosing their wealth. In this paper, we study the general case of the Millionaires’ Problem, in which a group of users try to compute the greatest value among their private inputs. We propose two solutions to address this problem. The first solution, which we call Smax-SH, is based on the AV-net protocol. It computes the greatest value while preserving the private input privacy in the semi-honest model. The second solution, called Smax-M, computes the maximum private value in the malicious model. The Smax-M protocol applies a zero-knowledge proof for security from malicious participants and active adversaries. We discuss the performance and security analysis of the proposed protocols and show that the each is efficient in terms of computation and communication costs. We also show that the Smax-M protocol is secure against a partial collusion attack in a malicious model.