| With the widespread use of mobile devices in people’s daily lives,mobile data traffic is experiencing explosive growth on the Internet.The traditional cloud computing paradigm is hard to support mobile services or applications with high-speed growth of large-capacity and diversified data,such as virtual/augmented reality,face recognition,and so on.In response to this challenge,Multiaccess Edge Computing(MEC),a new computing paradigm characterized by ultra-low latency,has emerged.This paradigm achieves this by deploying universal edge servers near mobile users at base stations or access points.Data caching is a universal technology that allows copies of data to be stored somewhere for quick retrieval when requested in the future.As a hot topic in edge computing networks,caching mainly includes two parts: data content caching and service caching.Appropriate data content caching decisions can improve user experience and increase service provider revenue;reasonable service caching decisions can improve task execution efficiency and reduce delays and energy consumption.Therefore,it is important to study caching strategies in edge computing.On the other hand,service caching and task offloading processes are closely coupled,and the cache decision and offload decision are constrained by the storage and computing resources of the edge server,which poses a challenge to the design of the cache decision and task offload strategy in edge computing.This paper mainly studies the data content caching problem and the task offloading problem under service caching in an edge computing system composed of multiple users and multiple edge servers.First,the data caching based on Popular Content Prediction in Multi-access Edge Computing problem under multi-access edge computing scenarios is studied,and a data caching revenue model is constructed.Under the constraint of storage capacity,with the goal of maximizing the benefit of the service provider’s data cache.Aiming at this goal,a DCPA algorithm based on Fully Polynomial Time Approximation Scheme is proposed.This paper verifies the performance of the algorithm through simulation experiments.The experimental results show that,compared with the comparison algorithm,the DCPA algorithm has better running time stability in large-scale scenarios,and the data cache revenue has increased by at least 33%.Second,this paper studies the Energy-aware Task Offloading in Service Caching(ETOSC).In this problem,the task offloading strategy is not only constrained by the computing resources of the edge server,but also affected by the location of the server and the attributes of the task itself.We prove that the ETOSC problem is NP-hard by reducing it to the capacitated facility location problem(CFLP),so the ETOSC problem cannot find an optimal solution in polynomial time.Therefore,the paper considers using methods such as linear relaxation to transform the problem into a submodular maximization problem,and proposes the ETOSCA algorithm by combining the solution of submodular maximization.Through simulation experiments,it can be found that compared with the comparison algorithm,the ETOSCA algorithm can save at least 8% of the energy cost. |
