移动社交网络中面向隐私保护的精确好友匹配

Privacy-preserving precise profile matching in mobile social network

好友匹配通过比较用户间属性相似度向交友请求者推荐好友,是移动社交网络应用中的核心功能。然而,在好友匹配的过程中,用户个人信息很可能会被服务器或者其他恶意用户窃取导致隐私泄露,且现有方案存在匹配结果不精确或者无法满足用户多维度隐私保护需求等挑战。基于此,提出面向隐私保护的精确好友匹配(P3M)方案,查询者可以根据自身需求灵活设定特征属性和距离的匹配范围。利用可比较内积编码(CIPE)和Paillier加密算法对用户属性和查询范围进行编码和加密,并设计安全点积协议实现用户属性和查询范围的安全比较。相较于现有方案,P3M方案支持查询者自定义查询范围以获得精确的查询结果,综合考虑用户特征属性及位置属性等多维度的隐私保护。最后,对P3M的正确性和安全性进行详细分析和证明,并通过实验验证P3M方案的有效性和高效性。

Profile matching is a key feature in mobile social networking applications,where friends are recommended to requesters by comparing the similarity of attributes between them.However,users’personal information is exposed to the risk of privacy disclosure in the process of profile matching.The existing solutions exist some issues such as inaccurate matching results or inability to meet users’requirements for multi-dimensional privacy protection.Baesed on this,a pri-vacy-preserving precise profile matching(P3M)scheme was proposed,which allowed users to flexibly set the matching range of attributes and distances according to their requirements.The Paillier encryption was utilized to ensure data secu-rity of users,and a secure dot product protocol was designed to achieve secure ciphertext comparison of user attributes and query ranges.The P3M realized multi-dimensional privacy-preserving of users including user feature attributes and location attributes.Finally,the correctness and security of P3M scheme were analyzed and proved in detail,and extensive experimental results verified the effectiveness and efficiency of P3M scheme.