Privacy leakage in two-party group nearest neighbor queries

Abstract

Location-based social network (LBSN) requires answering a variety of queries related to user locations. In this paper, we investigate issues in performing secure multi-party computation in LBNS and discuss some of the challenges and problems associated with it. In particular, we look at the group nearest neighbor queries and discuss the privacy leakage problem in two party group nearest neighbor computation. We elucidate the attack model and propose a privacy leakage index to measure the impact. We perform experiments to show the privacy leakage problem and suggest a solution to deal with this problem.