Research On Key Technologies Of Service Deployment For Edge Computing In Internet Of Vehicle

With the rapid development of wireless communication technology,the Internet of Vehicles(IoV)has given rise to many latency-sensitive services,such as intelligent driving and in-vehicle entertainment.Cloud Computing(CC)is limited by remote data transmission and the centralized architecture,which makes it difficult to execute services in real-time.Edge computing(EC)is a distributed computing paradigm that meets the service requirements of users by deploying edge servers(ESs)on the edge of the network,enhancing data processing capability and response speed,and reducing transmission latency and bandwidth consumption,thus effectively relieving the computation pressure on the cloud data center.However,due to the limited resources of ESs,it is difficult to balance the service latency and deployment cost when deploying services.Besides,the location of ESs will affect the performance of the EC system.The edge server placement(ESP)strategy with low performance affects the efficiency of service deployment,which makes it difficult to meet the service requirements of users and balance the service latency and workload of ESs.Therefore,the research of the service deployment for EC in IoV faces two challenges:(1)For the ESP,how to analyze the game relationship by the positions of ESs and optimize jointly the service latency and the workload of ESs.(2)For the service deployment,how to consider the heterogeneity of the IoV services and ESs to jointly optimize the service latency and deployment cost.Therefore,this thesis will study the service deployment for EC in IoV.Specifically,the research contents are as follows:(1)To solve the ESP for EC in IoV,an ESP method for joint optimization of the service latency and the workload of ESs is proposed.Specifically,this thesis first analyzes the ESP,constructs the transmission delay model,computation delay model,and workload model,and defines the ESP problem as a multi-objective optimization problem.Secondly,a noncooperative game model based on the locations of ESs is constructed.The optimal ESP strategy is solved by defining the benefit function and the updated rules of ESP strategies.Then the ESP method with the non-cooperative game(ESPN)is proposed.Finally,the convergence of the proposed method is analyzed.(2)To solve the service deployment for EC in IoV,a heterogeneous service deployment method for jointly optimizing service latency and deployment cost is proposed.Specifically,the heterogeneity of the service and ESs in IoV are considered first.Besides,the communication model,the computation model,and the heterogeneous service deployment cost model are constructed.Next,the heterogeneous service deployment problem is defined as a multiobjective optimization problem.To reduce communication latency and improve the efficiency of service deployment,ESs are divided into clusters based on distance and workload.Then,according to the principle of service type priority and the least number of the deployed service priority,the heterogeneous services are deployed in ESs.Next,the ES that handles service requests is selected based on the Analytic Hierarchy Process(AHP).Finally,a heterogeneous service placement method with AHP strategy selection(AHSP)is proposed.