SAFE: A social based updatable filtering protocol with privacy-preserving in mobile social networks

Mobile Social Networks (MSN), as an emerging social networking platform, facilitates social interaction and information sharing among users in the proximity. Spam filtering protocols are extremely important to reduce communication and storage overhead when many spam packets without specific destinations are diffused in MSNs. In this paper, we propose an effective social based updatable filtering protocol (SAFE) with privacy preservation in MSNs. Specifically, we firstly construct a filter Hash tree based on the properties of Merkle tree. Then, we exploit social relationships, and select those users with more than a specific number of common attributes with the filter creator. The selected users are able to store filters in order to block spams or relay regular packets. Furthermore, we develop a cryptographic filtering scheme without disclosing the creator´s private information or interests. In addition, we propose a filter update mechanism to allow users to update their distributed filters in time. The security analysis demonstrates that the SAFE can protect user´s private information from filter´s disclosure to other users and resist filter forgery attack. Through extensive trace-driven simulations, we show that the SAFE is effective and efficient to filter spam packets in terms of delivery ratio, average delay, and communication overhead.