SE-P²μm: Secure and Efficient Privacy-Preserving User-Profile Matching Protocol for Social Networks

By matching attributes in user profiles, social networking services (SNSs) are able to create certain social relationships between different users. However, user profiles often contain sensitive information about users. If data matching is done in plaintext, users’ privacy and security will face serious challenges. Most recently, the hacking incident of Ashley Madison website, which incurred more than 60 GB of data being leaked, has prompted researchers to explore the privacy of user data in social networks urgently. Existing schemes may be able to achieve privacy-preserving user profile matching, but they require high communication bandwidth and complex computational costs. To address them, we proposed a private user profile matching protocol for social networks using
$t$
-out-of-
$n$
matching servers. Applying the Chinese remainder theorem (CRT), Bloom filter technique, and skyline computation idea, the matching users are efficiently returned with the help of any
$t$
matching server without disclosing their identities. In addition, our scheme allows up to
$t$
matching servers to collude while simultaneously ensuring user profile privacy and query privacy. Finally, compared with two recent user profile matching protocols, the performance results demonstrate that our scheme outperforms them.