Research On Cache Strategy Based On Mobile Edge Network

With the rapid development of the mobile Internet,there is a huge demand for user services based on video and other content.At the same time,the accelerating growth of mobile devices means that more and more users are accessing the network to obtain content.However,actually,network resources such as bandwidth are limited,which makes it difficult for users to fully meet the low-latency and high-bandwidth content transmission requirements.How to solve the contradiction between user needs and limited network resources in mobile networks has undoubtedly become a big challenge.As an emerging caching framework,mobile edge caching caches highly popular content to network edge nodes with certain storage capabilities.A large number of studies have proved that content caching on edge nodes is one of the most effective technologies to reduce network traffic and overcome the bottleneck of centralized servers.When a large number of users request the same content,edge nodes can provide content transmission services and effectively preventing the core network from transmitting duplicate content and wasting bandwidth and other network resources to ensuring content service quality,which can provide users with a high-bandwidth,low-latency service environment.First,we studied the content caching strategy among multiple SBSs(Small Base Stations)in a cooperative wireless network.In current research results,the cache strategy is often designed around a single base station,and there is no cache-aware placement mechanism.In order to solve this problem,in the cache placement phase,we studied the mobile edge cache content placement strategy based on whether the system knows the prior knowledge of the requested file to maximize the cache hit rate.Specifically,for the case where SBS has prior knowledge of the requested file,we solve this problem through a caching strategy based on the KNAPSACK algorithm.For SBS without prior knowledge,we constructed a Markov Decision Process(MDP,Markov Decision Process)framework and proposed a Q-learning-based caching strategy.When the system observes whether the requested file is hit,the cache system observes the cache hit according to the files cached in the local SBS and its associated SBSSecondly,for user devices that also have caching capabilities,we combine the cellular network with the D2D(device to device)communication network,and use end user caching to increase throughput.At present,the existing research mainly discusses the cache between the base station and MBS(Macro Base Station,macro base station).Few literature studies the user cache and the correlation comparison between the user cache and the base station cache.Based on this,this paper considers that both the user equipment and the base station have the ability to cache content and can communicate with each other to improve QoS(Quality of Service).We studied the caching strategy under limited storage capacity in D2D cellular networks,and at the same time,considering the popularity of the content,we proposed a cache replacement strategy based on a heuristic algorithmFinally,in a mobile cellular network scenario,simulation is performed by selecting appropriate experimental data and system parameters,compared with traditional LFU(Least Frequently Used),LRU(Least Recently Used),FIFO(First in First out)and Greedy caching strategy in terms of performance indicators such as hit rate.The experimental results show that the KNAPSACK and Q-learning cache placement strategies proposed in this paper can improve the cache hit rate.In addition,in the D2D cellular network scenario,by comparing the MLPLRU(Multi Level Pareto Least Recently Used)algorithm and LRU,it is proved that the heuristic cache algorithm based cache replacement strategy proposed in this paper can effectively improve the cache hit rate.