Match-MORE: An efficient private matching scheme using friends-of-friends' recommendation

Abstract

Although Proximity-based Mobile Social Networks (PMSNs) bring mobile users novel ways to discover their similarities, they enjoy this kind of conveniences at the cost of user privacy and system overhead etc. To address these problems, we propose Match-MORE, which employs the concept of friends-of-friends to find some ones in common from friends and thus design a private matching scheme. Specifically, Match-MORE exploits a novel similarity function with considering social strength between users and similarity coefficient of the corresponding profiles, simultaneously. It provides users more opportunities to know other potential friends based on the recommendations from existing friends with tunable accuracy, and without disclosing too much private information to each other. The Bloom filter-based common-attributes estimation reduces the system overhead significantly. The security and performance are thoroughly analyzed and evaluated via detailed simulations.