Research On Edge Cooperative Cache Strategy Based On Local Content Popularity

Recently,with the number of the intelligent devices in the access network increase sharply,the huge data flow has brought severe challenges to the existing communication system.To deal with the problem caused by the data traffic,edge caching technology become research hotspots.The edge cache technology takes full advantage of the storage and computing resources of the edge nodes,and cache the content with higher popularity in advance,which brings the edge nodes closer to the user,thereby the pressure on the core network can be relieved.It exists three issues in formulating edge caching strategy,the prediction of the popular content,the placement strategy of cache content set,the update of cache content set.The article mainly focuses on clustering between base stations,popularity prediction based on user preferences,and content placement in collaboration domains.??In the 5G ultra-dense network(UDN),due to the less effective on base station clustering of traditional methods,which leads to the a lower cooperative cache performance,this article proposed a strategy of partitioned cooperative edge caching,by clustering the small base stations under the 5G ultra-dense network,many cooperative domains is obtained,and the cooperation between small base stations improves the overall cache hit rate.Firstly,according to the download history and location information between the small base stations,the cosine similarity and the Gaussian similarity between the small base stations are calculated,and the comprehensive similarity matrix is obtained.Then,the Fast Newman algorithm is used to divide the collaboration domain reasonably.Finally,according to the content popularity distributions in each domain,the content with a higher popularity is selected to cache in the local base station,other small base stations in domain and macro base stations hierarchically.The experimental results show that clustering small base stations through the modified algorithm can effectively divide the cooperation domain,and the overall hit rate can be increased with a lower average content transmission delay.In addition,the network performance is enhanced.??In each collaborative domain,with the goal of minimizing energy consumption,the optimal content placement is performed for each collaborative domain.Firstly,the global content popularity is calculated based on the global download history,and the Movie Lens data set is used to calculate the regional user preferences.By integrating the global video popularity information and regional user preferences,content popularity distribution is obtained.Then,define the content cache value,sort the cached content set according to the cache value of every content,and select the content with the higher cache value to cache.Finally,in the content placement strategy,considering the user request content preferences,base station location information,and the difference of energy consumption in each base station,a minimum energy consumption content placement strategy based on the greedy idea is proposed.The simulation results prove that compared with the most popular cache strategy and the best redundant cache strategy,the proposed strategy has a certain improvement in the performance of request hit rate,cache energy consumption and energy saving rate.In conclusion,the problems of base station cooperative caching and content placement in edge caching is solved in the article.Compared with the existing solutions,the network performance has been improved.