Gmatch: Secure and privacy-preserving group matching in social networks

Groups are becoming one of the most compelling features in both online social networks and Twitter-like micro-blogging services. A stranger outside of an existing group may have the need to find out more information about attributes of current members in the group, in order to make a decision to join. However, in many cases, attributes of both group members and the stranger need to be kept private and should not be revealed to others, as they may contain sensitive and personal information. How can we find out matching information exists between the stranger and members of the group, based on their attributes that are not to be disclosed? In this paper, we present a new group matching mechanism, by taking advantage private set intersection and ring signatures. With our scheme, a stranger is able to collect correct group matching information while sensitive information of the stranger and group members are not disclosed. Finally, we propose to use batch verification to significantly improve the performance of the matching process.