| In recent years,with the development of 5G and the surge in demand for low latency services,the European Telecommunications Standards Institute upgraded mobile edge computing to multi-access edge computing for the development needs of 5G in 2017.Multi-access edge computing is to deploy the server at the edge of the wireless access network and provide computing,storage and intelligent service API capabilities nearby.However,due to the limited computing power,storage resources and the limited network coverage of the edge server,the mobility of users may deteriorate the service quality or even interrupt the service,thus affecting the user experience.As a way to compensate for the lack of mobility of edge computing,service migration can maintain the continuity and quality of service.In previous research on service migration,most of them are based on 4G network architecture for algorithm design.However,with the development of 5G,how to design service migration algorithm based on 5G architecture also needs further discussion and research.This paper studies how to improve user experience by improving service migration performance based on the separation of control plane and user plane in 5G network architecture.(1)This paper analyzes the relevant theories and research status about 5G user plane and service migration technology,and then proposes a heuristic UPF deployment algorithm based on region division.The algorithm considers the service request volume and UPF capacity,and aims at minimizing the average network delay and UPF deployment cost.Based on an improved heuristic algorithm,it proposes a UPF deployment scheme in the current mobile edge computing environment.The simulation results show that the algorithm can propose a reasonable UPF deployment scheme to reduce the average network delay of users and the UPF deployment cost.(2)Based on the characteristics of UPF distributed deployment,a new service migration model is further proposed.An online service migration algorithm based on Q-Learning is proposed to solve the problem of long-term delay and cost optimization in the single-user service migration model.The simulation results show that the algorithm can effectively reduce the UPF transmission delay of users,and optimize the total delay and total cost of users in the whole mobile process. |
PDF Full Text Request