Research On Edge Collaborative Caching Algorithm For Mobile Video Service

In recent years,the rapid development of mobile Internet is greatly changing the living habits of mobile users from all aspects.Mobile video service has increasingly become the main network application of users.However,mobile video service generates massive mobile network traffic,which brings a huge burden to the infrastructure of mobile network.The massive traffic has a bad influence on the performance of mobile network and deteriorates user's experience.Solving this problem requires not only better network infrastructures,but also an optimization from the perspective of data traffic.Mobile edge caching technology is an important part of the Mobile Edge Computing.Caching servers with large capacity are deployed in the mobile base station to cache the popular mobile video contents frequently requested by users,which provides users with faster services.However,compared with the total amount of services,the available resources of mobile edge network are relatively limited.Limited caching space restricts its service ability.The rapid development of user intelligent terminals not only brings new mobile applications,but also has an important impact on the structure of mobile network.These intelligent terminals have become a potential network resources to be utilized.In order to make better use of the cache resources of the edge network and provide high quality service to users,it is necessary to study the cache strategy of mobile edge cache.Based on the scenario of the mobile video,this thesis studies the edge caching scheme of the cooperation between edge and terminal devices.Aiming at the problem of initialization of edge cache content,this thesis proposes an active caching method for mobile video service with multi-dimensional service request adaptation.This thesis analyzed and predicted the service request behavior of mobile video service users.Then the D2D-assisted collaborative active caching is studied and the optimization problem with hit rate as the target is described.This thesis designs a replica repeated auction algorithm to solve the content placement problem.The simulation results show that the proposed algorithm achieves a higher cache hit rate.It means that the algorithm can make better use of the cache space on the edge and improve the service ability of the edge cache.Aiming at the problem of contents updating and replacing,this thesis proposes an intelligent replacement method of cached contents which minimize the service delay.Taking the delay of the user playing mobile video as the target,the user-base station collaboration in the mobile network is modeled.The analytical expression of the delay is obtained.To solve the cache content replacement problem.The problem is described as a Markov decision process.And this thesis designs a cache replacement algorithm based on deep Q learning.The simulation results show that the proposed algorithm achieves lower access delay and higher cache hit rate and reduces the traffic burden from the core network.In summary,this thesis studies two methods based on user-base station collaborative edge caching for mobile video.Users' requests can be satisfied in the edge network through the collaborative edge cache.Collaborative edge cache on the one hand relieves the pressure on the backhaul link and the core network,on the other hand reduces the service delay and improves the user experience.