Research On Edge Network Content Caching Model And Multilayer Strategy

In the 5G era,mobile devices and wireless network services are rapidly developing,and data traffic is showing explosive growth.Users have high requirements for real-time services such as social media,high-definition video conferencing,and live streaming.In this case,caching content in edge network systems has received widespread attention in order to meet users’ high-quality requirements.Due to the limited cache resources of devices in the edge network,it is difficult to maximize the effective use of cache resources in network devices,reduce Transmission delay,and improve cache hit ratio in the edge cache.In order to solve the above problems,this thesis focuses on predicting content popularity,improving cache hit rate and reducing Transmission delay.The main work of this paper is as follows:Firstly,this thesis proposes two types of content caching system models in edge networks.They are:(1)In order to improve the cache utilization of edge networks,an edge collaborative caching model based on content popularity is proposed.Consider the popularity of content,the cache capacity of edge servers,and the distribution of locations between them to establish the problem of maximizing cache hit rates.(2)User devices share content through D2 D in the user collaboration domain;The edge servers are divided into multiple clusters using the K-Means algorithm,and the edge servers between the clusters provide content to users in a collaborative manner.Then,we analyze the Transmission delay generated by the two-layer collaboration request content to establish the content cache problem,and propose a two-layer collaboration edge cache model based on delay and cost balance.Secondly,based on the proposed system model,this thesis designs two content caching strategies.(1)We propose an edge collaborative caching strategy based on content popularity prediction,which predicts user preferences through the Labeled LDA model.Then,we combine the probability of user requests for content with user preferences to achieve content popularity prediction.Finally,a content caching algorithm based on Q-learning is proposed to solve and obtain the optimal content caching matrix.(2)This thesis proposes a double-layer collaboration edge cache strategy based on delay and cost balance,analyzes the Transmission delay generated by the content of double-layer collaboration requests and the cost caused by the cache content,to establish the phase equilibrium problem between the delay gain and the cache cost gain,and proposes a delay cost balance algorithm to solve it to obtain the optimal content cache.In addition,a Markov model is constructed to predict content popularity,making it constantly changing over time.Through simulation verification,the edge collaboration caching strategy based on content popularity prediction in this thesis is compared with other caching strategies that consider cache hit rates,verifying that this algorithm has better cache hit rates.The double-layer cooperative edge caching strategy based on delay and cost balance in this paper is compared with other caching strategies considering Transmission delay,which verifies that the caching strategy in this paper effectively reduces Transmission delay.