Research On Server Deployment For Mobile Edge Computing

The development and appliance of cloud computing technology has greatly eased the computing and storage pressure of mobile terminals,and has provided important support for the diversity of mobile services.However,although centralized cloud computing centers can provide powerful computing and storage capabilities,long-distance transmission between mobile terminals and cloud computing centers may cause significant delays to delay-sensitive services.In this case,mobile edge computing technology is effective by sinking and distributing the computing and storage resources of the cloud computing center to the edge of the network,shortening the propagation distance of business data,and solving the delay problem of mobile terminals initiating and running cloud services,so as to effectively improve the quality of experience of mobile users.In this thesis,the problem of edge server deployment is mainly analyzed.Firstly,the background and main research issues of mobile edge computing are introduced,and then it is proposed that the goal of the research in this thesis is to deploy edge servers,and analyze the common deployment strategies of edge servers in combination with related works.Then,the four kinds of architecture of mobile edge network and the common server deployment scenarios and edge computing application scenarios are introduced,and the deployment requirements and common deployment methods of edge servers are briefly described.Finally,an optimized deployment scheme of edge server is designed and proposed.This scheme initially divides all base stations in the region according to neighboring density under the condition of satisfying the delay constraint and load constraint,and further optimizes the initial solution set on this basis.Through base station sequencing disturbance,server location offset,and server number reduction to perform multi-level clustering iterations to continuously optimize the number and deployment location of servers.And then the K-means Clustering deployment scheme and the Random deployment scheme are used as the comparison schemes of the proposed scheme to compare the performance on three aspects of average access delay,user experience stability,and load balance.The simulation is based on the Matlab platform.The experimental results show that the multilevel clustering iterative deployment scheme proposed in this thesis is superior to the other two deployment schemes in performance,which effectively improves the server load balance and the stability of user service experience.