Research On Edge Server Dynamic Selection Method In Mobile Environment

As a new computing mode,edge computing can help user complete complex tasks with limited battery resources and computing power.The users can directly connect with the edge server deployed nearby and obtain the required resources,avoiding the data transmission among network nodes and reducing the waiting time of users.However,it has its own limitations when the users invoke services deployed in edge server.On the one hand,the resources of the edge server are limited.If there are too many users connecting to the same edge server in a period of time,it may cause overload of the server,and be unable to provide users with the required resources,as well seriously affect the user experience.On the other hand,the edge server is usually deployed with the base station(or wireless access point),so that the coverage of the edge server is limited.Only the users within the edge server's coverage can connect to the edge server.However,in real life,the users often move,and need to switch to different edge servers to connect in the process of moving.If there are still unfinished tasks on the original server in the process of server switching,migration should be considered,and the delay caused by service migration has a serious impact on user experience.Therefore,this paper focuses on how to dynamically select the appropriate edge servers for mobile users to connect.We first study the way that dynamically selects edge servers for single moving user.And based on this,we study how to dynamically select edge servers for multiple users in mobile environments.The contributions of this research are as follows:(1)In the process of dynamically selecting edge servers for single moving user,the resource requests of other users to the edge server are unknown,and the limited resources of the edge server may not meet the all the resource request.Therefore,we propose the task completion probability model of the edge server according to the resources of the edge server,and analyze the calculation cost in the two cases of both service migration and non-migration.And then combining the advantages of genetic algorithm and simulated annealing algorithm,an edge server selection algorithm for mobile single user is proposed,called GASS(combined Genetic algorithm and simulated Annealing algorithm for edge Server Selection).In the mobile environment,the user connects to the edge server selected by GASS algorithm to optimize the user's time delay and energy consumption.(2)In the process of dynamically selecting edge servers for multiple mobile users,we build a model to calculate the total waiting time of multiple users.Meanwhile,we analyze the resource competition among multiple mobile users in mobile environment,and design an algorithm MESP(Multi-user Edge server Selection method based on the Particle swarm optimization)to select edge server for multiple users.In polynomial time,the MESP algorithm dynamically selects the connected edge servers for all users to minimize the total waiting time of all users.(3)In this paper,a random path is selected from Baidu map as the known mobile path of users,and a simulation experiment was designed based on the data set of the location of Shanghai Telecom base station.In the experiment of dynamically selecting edge server for single user,the effectiveness of GASS algorithm is proved by comparing genetic algorithm,simulated annealing algorithm,greedy algorithm and random algorithm,and the influence of user's moving speed and edge server coverage radius on experimental results was analyzed.In the scenario of dynamically selecting edge server for multiple mobile users,we take greedy algorithm and random algorithm as baseline approaches,verified the effectiveness of the proposed MESP algorithm,and analyzed the impact of different data round-trip delay and service migration on the total waiting time of all users.