| With the development of technologies such as artificial intelligence,augmented/virtual reality,and 5G,the demand for high dynamic real-time applications such as assisted driving and intelligent inspection,which have strong user mobility and high variability in demand,is gradually increasing for network bandwidth and computing efficiency.Traditional centralized cloud processing solutions will bring significant transmission latency and computing pressure,which cannot meet the business requirements in highspeed moving scenarios.The integration of Mobile Edge Computing(MEC)and 5G has become a hot issue in academia and industry,with the key problem to be solved being how to efficiently achieve collaborative work among edge nodes and ensure the reliability of services.In the existing edge service network,edge nodes differ greatly in geographical location,computing power,and cache capacity,and the user’s service requirements have significant time-varying characteristics,making it difficult to dynamically evaluate and aggregate the service capabilities of edge nodes.Existing edge service network solutions mostly consider point-to-point edge node collaboration models,which do not effectively aggregate edge resources,making it difficult to guarantee the continuity and reliability of edge services.On the other hand,existing edge computing application models mostly focus on scheduling business or traffic,without detailed differentiation of the computational requirements of various service functions in the business.For example,service components(i.e.,software)are an indispensable basis for performing computational tasks,but the existing edge computing service models have not evaluated and optimized the service capabilities of edge nodes fr-om the perspective of service components.Edge nodes can cache the required service content locally,effectively reducing transmission latency.However,in existing content caching solutions,edge nodes mostly download content data from the cloud platform based on business computational requirements during the service process,without fully considering pre-caching based on user behavior characteristics,resulting in higher service response latency and lower content delivery success rate.To address the problem of low resource efficiency caused by difficult user demand prediction and insufficient aggregation of multi-edge node service capabilities,this paper proposes a dynamic construction method for the edge collaboration service domain by clustering multiple edge nodes and coordinating the sharing of service components and collaborative delivery among edge nodes by the domain head node,achieving a balance between collaborative domain service capabilities and stability.In the constructed edge collaboration service domain,to further solve the problem of poor service quality caused by unreasonable content configuration,a utility-balanced edge collaboration service component pre-caching mechanism is proposed.The domain head node can formulate service component caching decisions based on user preferences and cache feedback,thereby balancing and improving network service capabilities,cache accuracy,and resource utilization.To address the problem of low delivery success rate of service components caused by user high-speed movement,a multi-mode delivery mechanism for service components based on user behavior prediction is proposed,which allows edge nodes to efficiently and accurately deliver cached service components to users based on their location and the distribution of required service components,through intra-domain node collaboration,inter-domain collaboration,and cloud-edge collaboration modes.A collaborative domain service component update mechanism based on popularity evaluation is designed to update and replace the stored service components in edge nodes according to the future popularity of service components,to meet the time-varying service requirements of users.Simulation results show that the proposed edge collaboration service component caching and delivery mechanism based on edge collaboration domains can effectively reduce the service response latency of edge nodes,improve service quality,and increase resource utilization efficiency. |