The Research Of Privacy-preserving Friend-finding For Multi-relational Social Networks

A boom in mobile hand-held mobile devices and mobile applications has made Mobile Social Networks(MSNs) experience an explosive growth recently. MSN enable people to access existing online social services. Meanwhile, numerous mobility-oriented applications, such as location-based services are introduced. As one of the most popular applications in MSN, friend matching allow users to make friends with those who share common attributes based on uses’ profiles. The key challenge of friend matching is how to protect users’ privacy during the matchmaking process. In this thesis, solutions based on private set intersection are provided to solve the privacy-preserving matchmaking problem under different circumstances. The main contributions are the following:1. The centralized system architecture is proposed, under which a two-phase privacy-preserving matchmaking protocol against malicious users is designed. The theoretical analysis of our protocol about privacy level, efficiency and security is performed. The simulation experiments are also performed to compare the efficiency of our matchmaking protocol with existed ones.2. Under distributed circumstances, we propose two matchmaking schemes. The correctness, security and efficiency of two proposals are analyzed and the experiments are carried out to verify the efficiency.3. Based on the distributed protocol, a fast privacy-preserving friend discovery scheme is designed. We perform the experiments to study the cost and impact of the friend discovery scheme under different scale networks.