Trust evaluation based friend recommendation in proximity based mobile social network

Abstract

Proximity Based Mobile Social Networks (PMSN) is a special type of Mobile Social Network (MSN) where a user interacts with other users present in near proximity in order to make social relationships between them. The users’ mobile devices directly communicate with each other with the help of Bluetooth/Wi-Fi interfaces. The possible presence of a malicious user in the near proximity poses significant threats to the privacy of legitimate users. Therefore, an efficient trust evaluation mechanism is necessary that enables a legitimate user to evaluate and decide about the trustworthiness of other user in order to make friendship. This paper proposes a protocol that evaluates various parameters in order to calculate a trust value that assists a PMSN user in decision making for building new social ties. This paper describes various components of trust such as friend of friend, credibility and the type of social spot where trust evaluation is being performed and then calculates these parameters in order to compute the final trust value. We utilize semi-trusted servers to authenticate users as well as to perform revocation of malicious users. In case of an attack on servers, real identities of users remain secure. We not only hide the real identity of the user from other users but also from the semi-trusted servers. The inherent mechanism of our trust evaluation protocol provides resilience against Sybil and Man-In- The-Middle (MITM) attacks. Towards the end of the paper, we present various attack scenarios to find the effectiveness and robustness of our trust evaluation protocol.