Privacy-preserving profile matching protocol considering conditions

Abstract

A social matching service has recently become popular. These services help a user to search friends having common preference or interest. On the other hand, users use their personal information for matching in social matching services, and thus the privacy-preserving profile matching protocols have been well studied. However, although there are various privacy-preserving profile matching protocols, they may cause unwilling matching. In order to solve this problem, it is necessary to achieve a fine-grained matching mechanism considering conditions. In this paper, we propose a privacy-preserving profile matching protocol embedded with homomorphic encryption considering conditions: matching is established only when the conditions are satisfied. Our protocol reduces computational cost of user’s device by using the map-toprime technique and setting an honest-but-curious server. Furthermore, even if a server is attacked, user’s secret key or personal data does not leak since our protocol is designed for a server without such confidential data.