Virtual private social networks

Abstract

Social Networking Sites (SNSs) are having a significant impact on the social life of many people - even beyond the millions of people that use them directly. These websites usually allow users to present a profile of themselves through a long list of very detailed information. However, even when such SNSs have advanced privacy policies, users are often not aware of their settings and, on top of that, users cannot abstain from sharing a minimum set of information (e.g. name and location). Such a small set of information has been proven to be enough to completely re-identify a user [22, 25]. In this work we introduce the concept of Virtual Private Social Networks (VPSNs), a concept inspired by the one of Virtual Private Networks in traditional computer networks. We argue that VPSNs can mitigate the privacy issues of SNSs, building private social networks that leverage architecture publicly available for SNSs. Furthermore, we propose FaceVPSN, which is an implementation of VPSNs for Facebook, one of the most used SNSs. FaceVPSN is the first privacy threats mitigation solution that has a light and completely distributed architecture - no coordinator is required. Furthermore, it can be implemented without any particular collaboration from the SNS platform. Finally, experimental evaluation shows that FaceVPSN adds a limited overhead, which, we argue, is acceptable for the user.