Private and Flexible Proximity Detection in Mobile Social Networks
A privacy-aware proximity detection service determines if two mobile users are close to each other without requiring them to disclose their exact locations. Existing proposals for such services provide weak privacy, give low accuracy guarantees, incur high communication costs, or lack flexibility in user preferences. We address these shortcomings with a client-server solution for proximity detection, based on encrypted, multi-level partitions of the spatial domain. Our service notifies a user if any friend users enter the user’s specified area of interest, called the vicinity region. This region, in contrast to related work, can be of any shape and can be flexibly changed on the fly. Encryption and blind evaluation on the server ensures strong privacy, while low communication costs are achieved by an adaptive location-update policy. Experimental results show that the flexible functionality of the proposed solution is provided with low communication cost.