Research On Privacy-Preserving Technique Of Friend Discovery In Mobile Social Network

With the development of wireless communication technology, mobile positioning technology and smart phones, mobile social network has developed rapidly in recent years, and it has become an absolutely necessary part of people’s life. Privacy is the secret that people do not want to public and to tell anyone else. The problem of revealing privacy in the Internet has brought a lot of concerns in recent years. People enjoy the convenience of mobile social network, at the same time there is a serious problem of revealing privacy. Existing research has less attention on privacy-preserving in mobile social network, and there are some problems about revealing privacy information that need to be resolved, such as a user’s location information revealing, attributes information revealing and so on.Friend discovery means that finds one’s old friends or potential friends, and it is one of the major functions provided by the mobile social network. The thesis carried out a research on privacy-preserving of friend discovery in mobile social network, and proposes a solution for security in friend discovery. First of all, we compute the proximity between users and filter them after clustering them based on each one’s position. We design a tripartite dot product protocol based on paillier encryption algorithm, which compute the dot product values between two users based on the encrypted vectors of users’ privacy attributes, and then, we judge whether the dot product values meet the requirements of the certain threshold. In this process, the user’s privacy information and the dot product values are not be leaked to anyone else. Secondly, after filtering users based on their proximity, we introduce the concept of trust degree, and use the two attributes trust degree and proximity to achieve collaborative filtering of user groups. This thesis proposes a trust model based on friendship between users, which can be used to evaluate user’s trust degree in mobile social network. In the trust model, we use the reputation to represent the global trust degree, and the numbers of mutual friends between users represent the indirect trust degree. And then, we make the two quantitative data together to get the user’s trust degree. By ranking the trust degree, it can filter the attacker’s illusory nodes to some extent, and increase the using cost of the attacker.Finally, we demonstrate the designed dot product protocol and trust model with performance test on simulation experiments. Through analysis of experimental result and the protocol, it may draw the conclusion that the dot product protocol and trust model can protect the users’ privacy to some extent on the premise of implementing users’ requirements. And it reached the purpose that processing security friend discovery with privacy-preserving in mobile social network.