Research On Edge Server Deployment Based On K-means Algorithm

Cloud computing and mobile cloud computing have experienced rapid development.Although centralized cloud computing provides abundant computing resources for intensive tasks,the instability between intelligent terminal devices and cloud increases the difficulty of handling sensitive data,and there will be unpredictable network latency.In order to solve this problem,the mobile edge computing is proposed and the cloud computing is pushed closer to the edge in the intelligent terminal network.The edge servers are deployed at the network edge,providing low latency and high bandwidth services and improving the quality of service for intelligent terminals.In this paper,the problem of the deployment for edge servers in the mobile edge computing environment is studied and analyzed further.IOT(Internet of things)devices locally producing massive data that stored in the smart terminal deploy both the optimal location and the optimal number of edge servers in an efficient and economical way and the relevance of the intelligent terminals and edge servers is also considered so the maximum number of intelligent terminals and optimal quality of service can be made in the same time.Based on the K-means clustering algorithm,we deploy the edge servers reasonably and compare the influence of Randomly-deployed algorithm,Density clustering deployment algorithm and K-means clustering deployment algorithm on the average completion time of the system.In the experimental section,the trade-off for the time revenue and the deployment cost gain is analyzed and the optimal number of deployed edge servers is found;at the same time,the influence of network delay threshold on three algorithms is analyzed.The experiment is based on the Matlab R2016 b platform.The simulation results show that the K-means clustering deployment algorithm is better that has been verified by changing the number of intelligent terminals.