Research On Neighbor Discovery And Friend Matching In Proximity-based Mobile Social Networks

With the rapid development of mobile networking technology and mobile terminal device, more and more people are now using different kinds of mobile social network services. User interaction can be initiated by neighboring mobile users via Bluetooth or WiFi of their smart terminal devices directly without having to access internet, creating a proximity-based mobile social network, which is convenient social network mode for people to make new friends and communicate face to face, and it has been drawing more and more attention. Neighbor discovery and friend matching are the foundation of how the proximity-based mobile social network works. Research in finding highly efficient solution is of great significance for the promotion of the theoretical study and practical application of this kind of social network.Firstly, a Bi-direction-based method of neighboring-user-detection is proposed. A static active-period slot and two dynamic active-period slots are used in the method. Two dynamic active-period slots move oppositely from the right and the left in every cycle. Fringe detection principle is adopted in Bi-direction for non-aligned mode, which greatly reduces the duty cycle needed by neighboring-user-detection and enables fringe detection to serve the aligned mode. Overflowing fringe detection method is designed and the idea is proposed which suggests that randomly selecting the open position of dynamic active-period slots for the first time may further improve the detection efficiency.Secondly, a neighbor-acceleration-based method of neighbor discovery is proposed. The idea of indirect neighbor discovery is adopted to add detected neighbors into the neighbor list, and select the position and quantity of extra opened active-period slots through the scheduling algorithm of active-period slots and the remaining power of the node to further improve the efficiency of indirect neighbor-detection. This method can be jointly used with any base method, which not only improves the performance but also makes it more suitable to be applied in a mobile environment.Thirdly, a distributed-computing-based method of dynamic friend-matching is proposed. Two levels of privacy protection are defined. User attribute priority can be dynamically adjusted according to user’s demand. Paillier Encryption is adopted to encrypt the high priority user attribute information, guaranteeing the safety of user information via its homogeneity and self-shielding. An early filtering agreement is designed to remove those candidate users who fail to meet the threshold conditions. The Privacy Protection of Level I is applied to the friends-matching through correlation coefficient method. The generalized Jaccard coefficient method is designed to apply the Privacy Protection of Level II for the matching agreement.Lastly, a dual-server-based third party friends-matching method is proposed. An anonymous server and a computing server are adopted at the same time for friend-matching in this method. The user information is divided into the user ID information and the priority information of user attribute, which are encrypted and transmitted with the public keys of anonymous servers and computing servers respectively. The replacement and anti-replacement treatment of the user ID information are achieved by anonymous servers, fulfilling the objective of sabotaging and recovering the corresponding relationship between the user ID information and the priority information of user attribute. The generalized Jaccard coefficient method is adopted by computing servers to compute the user similarity, yet it is unable to acquire the corresponding ID information of real users of the user attribute priority, which remedies the shortcomings of traditional third party computing method and realizes the friend-matching under the Privacy Protection of Level III.The experiment result shows that the methods of neighbor discovery proposed in the paper improve the efficiency of neighbor-detection and they are more suitable to be applied in a mobile environment. The distributed friends-matching method and the third party friends-making method not only greatly reduce the cost of computing and communication of mobile terminal devices, but they are also applicable to different contexts of use, which are helpful for the application and promotion of neighboring social network.