Research On Cooperative Content Caching And Delivery In A Cloud,Edge,and Fog Computing System

In recent years,edge computing technology represented by Mobile Edge Computing(MEC)and Fog Computing has been considered as one of the key technologies to alleviate network bandwidth pressure and to deliver services at low latencies.Compared with traditional data centers,edge computing nodes have the features of closing to end users,wide distribution;however,they are also suffered from the disadvantages of limited resources including storage and computing resources.Therefore,facing tremendous user requests and massive Internet data traffic,an efficient content caching and delivery strategy is key to making full use of edge nodes and improving user experience.This thesis just focuses on the content caching and delivery strategies based on the Cloud,MEC and Fog networks considered.We first introduce the architecture of Cloud-MEC,and based on this architecture,we propose a zone-based cooperative content caching and delivery strategy.Under this strategy,the MEC servers in the adjacent sub-zone cache content and provide services to users cooperatively.We formulate the optimal cache problem for each MEC server.We also develop a Mixed Integer Linear Programming(MILP)model for the problem and also propose a MixCo heuristic algorithm to decide the cached content in MEC server.The novelty of this algorithm is to divide the storage space in each MEC server into two parts.The first part is to cache the most popular content locally in the corresponding area served by the MEC server,and the second part will collaborate with other MEC servers to cache less popular content in the whole zone.The simulation and experimental results show that compared with the traditional strategy,this new proposed strategy can reduce the average content delivery latency and balance the loads on different MEC servers.We further look into content caching and delivery problem in the context of the Cloud-MEC-Fog architecture,where distributed Fog nodes can also provide their resources for the content storage and delivery purpose.We study the impact of different caching strategies deployed for the MEC and Fog nodes on user experience.For the Fog node,we propose a caching strategy,called the RMC(Remove Max Copy)strategy,which replaces contents with the largest numbers of copies in the Fog network at the first priority.We also develop a dynamic simulation model to simulate the processes of content release,content request,and popularity change in real world.The simulation results show that the combination of the MixCo and RMC strategies can greatly improve the hit ratio and enhance the user experience.