Fine-Grained Privacy-Preserving Spatiotemporal Matching in Mobile Social Networks

With the rapid popularization of mobile smartphone and its built-in location-aware devices, people are possible to establish trust relationships with each other by matching their interests, hobbies, experiences, or spatiotemporal profiles. However, the contradiction between the possibility of personal sensitive information leaking and the growing privacy concerns of users restricts the widespread use of direct matching schemes. To addressthis problem, lots of privacy-preserving matching schemes were proposed recently years. These schemes ensure users find the perfect matcher(s) without revealing extra unnecessary personal information. And yet, at the same time, it is inevitable that they produce more computation amount and communication traffic compare with former direct matching schemes. For mobile application scenarios, it is a heavy burden since power is limited. Particularly, for spatiotemporal matching, the situation is much worse due to the number of elements in users' spatiotemporal profiles will be very large as time goes on. Another outstanding issue in spatiotemporal matching is that how to define two users are neighboring. So, how to achieve an efficient and exactly privacy-preserving spatiotemporal matching remains an open question. In this paper, we propose a fine-grained privacypreserving spatiotemporal matching in Mobile Social Networks. Our scheme decreases the spatiotemporal matching error, as well as promotes the efficiency of matchmaking. Thorough security analysis and evaluation results indicate that our scheme is effective and efficient.