Research On Social Network Data Placement Strategy In Cloud-edge Computing

With the development of technology and the popularization of smart end devices,social network applications such as Weibo,YouTube,and Twitter provide people with convenient communication methods.Traditional social media mainly consist of pictures and articles content.In recent years,with the development of cloud computing,Internet of Things and other technologies,various new social network applications represented by new media contents such as interactive live broadcasting and real-time conferences have begun to appear in users' life.In such applications,users are more sensitive to access latency of various new media.At the same time,the popularity of smart end devices has brought hundreds of millions of users to use social networking applications for real-time communication,the amount of data generated by it also increases explosively.It is becoming more and more important to maintain a reasonable load balancing of the storage system and ensure good system performance.Through the cloud computing can provide better storage services for massive users' data in traditional social network applications platforms,cloud datacenters are generally far away from users,and it is difficult to ensure users' high real-time requirements for various types of new social media.With the emergence of edge computing,various computing and storage resources have to sink to edge servers that are closer to users.In view of the above problems,this thesis has done the following two aspects of research work based on Graph-Partitioning Algorithm(GP).Firstly,the social network data placement strategy in cloud computing is optimized in this thesis.With the constraints of users' access latency and load balancing,combining GP with data placement cost model,a Balanced Graph-Partitioning Algorithm(BGPA)is proposed in this thesis to optimize data placement cost while ensuring users' latency requirements and keeping load balancing of datacenters.Then the cost of social network data placement strategy in the cloud-edge computing is optimized in this thesis.Combination of the different access latency requirements of users,a Cost-Effective Different Latency and Load Balancing-Constrained Data Placement Strategy in Cloud-Edge Computing(CSCE)is proposed in this thesis,the cost of data placement is optimized under the condition of ensuring the user's access latency requirements and maintaining load balancing of cloud-edge computing.This thesis researches the cost optimization of social network data placement based on the two algorithms proposed above.This thesis not only optimizes the load balancing degree between datacenters in cloud computing but also optimizes the data placement cost while ensuring users' access latency lower than 200 ms.On this basis,the different latency requirements(200ms,10ms)of users for social content in cloud-edge computing are considered in this thesis,and the cost of data placement is optimized while maintaining system load balancing.In order to verify the effectiveness of the strategy,real Facebook data sets are used in this thesis for simulation experiments,and the experimental results prove the effectiveness of the algorithm proposed in this thesis.