Research On Data Diversion Based On Mobile Social Network

With the vigorous development of the Internet,the number of mobile devices in China is increasing,which obviously brings a huge traffic burden to mobile communication operators.At present,mobile social network is facing the problem of how to find a balance between relieving network pressure and meeting user needs.Therefore,data offloading emerges as a new technology of data processing,which has been widely studied in the academic field.The basic principle is that some users directly request data from cellular network,and then share data to neighboring users through short-distance communication by using the opportunity contact brought by their mobile,and play the role of data diversion.If the data is not received by the adjacent users after a certain time delay,the data can be obtained by means of cellular network.One of the key points of data offloading is how to select some users as the secondary data offloading of seed nodes.Therefore,the content of the paper is as follows:(1)This dissertation proposes a D2 D data offloading method,DDS(dynamic demand scenario),based on the dynamic needs of users in mobile social networks.Considering the user's interest and moving trajectory in different periods of the day,a graph structure is constructed to describe the user's traffic demand.Secondly,mobile users are divided into different communities according to their interests by Newman fast algorithm.Users in the same community have similar data needs and high contact frequency.Finally,the centrality measurement algorithm is used to obtain seed users in each community to maximize the data offloading of cellular network.In order to select the appropriate seed node,this dissertation compares five different centrality measurement methods.The empirical conclusion shows that in the DDS strategy,the seed node offloading effect based on Page Rank measurement method is the best.In this dissertation,DDS is designed and compared with other offloading methods to further verify the reliability of DDS scheme.The experiment proves the feasibility of the new DDS scheme.(2)This dissertation proposes a D2 D data offloading method based on user integration social relationship.Firstly,data content items and user information are established to represent different interest attributes of users.The user content rating matrix is established to represent the user's preference for certain data content.Secondly,in order to explore the social relationship of real users,extreme value theory is used to calculate a threshold.If the contact time of two users exceeds this threshold,they are considered to have social relationship;on the contrary,there is no social relationship.Based on this,a trust matrix is generated to indicate the user's social relationship.Finally,the trust and content score are integrated to divide the community and select the seed node.In the end,we propose TS(Trust-based scenario),PS(Preference-based scenario)and CS(Comprehensive-based scenario),and use the centrality measurement strategy to select the best seed user.Among them,the optimal seed user selected by CS strategy can share data with a large number of users,which greatly reduces the cellular network traffic and has a good data offloading effect.The new scheme fully considers the dynamic needs and social relationships of users.According to the comparative experiments,the effectiveness of the new scheme in data offloading is confirmed,which can alleviate the traffic load of cellular network.But the two methods do not fully consider the user's selfishness.Therefore,the future work will introduce incentive mechanism to stimulate users to actively participate in data offloading.