Research On Video Caching Strategies For Mobile Edge Network

With the rapid development of Internet communication and the wide application of intelligent terminals,video traffic is spiraling.Moving video cache to the edge of the network is an effective solution to unload core network traffic and shorten the delay of users accessing content.However,the existing cache work lacks the comprehensive consideration of users and video content,resulting in low cache hit rate and low accuracy of the whole system.Mobile edge network technology can use widely distributed mobile edge infrastructure to store and distribute videos,which can not only effectively improve the video distribution capacity and relieve the congestion problem of the core network,but also effectively assist users to obtain videos on the network near the user side and reduce the access pressure of remote video content servers.And according to the video popularity forecast and dynamic update edge server cache.Therefore,cache mechanism is one of the core technologies of mobile edge network,and the research hotspot in the future network field.In this paper,the decision strategy and replacement strategy of video cache in mobile edge network are studied respectively.(1)Aiming at the uneven distribution of users,the heterogeneity of user preferences and the difference of video content characteristics in mobile edge networks,a video caching decision strategy in mobile edge networks is proposed to improve the caching efficiency and prediction accuracy of video content.Firstly,a collaborative cache model for user request content and video content prediction is constructed,and users are clustered using a clustering algorithm based on Chichi information criterion;Then,according to the clustering results,the echo state network is used as the machine learning framework to predict the video;Finally,according to the prediction results,the video with high access probability in the future will be cached to the small base station.The simulation results show that compared with KNN,LCE and LCD cache decision strategies,this method has significantly improved cache hit rate,accuracy and recall rate.(2)Aiming at the problem of insufficient storage space of edge server in mobile edge network,a video cache replacement strategy in mobile edge network is proposed to achieve better cache hit rate.Firstly,a video prediction model based on mixed multiple regression is constructed,which considers many factors such as viewing time,sharing times and popularity of video content;Then,according to the predicted video popularity,the unpopular video content is dynamically replaced.This strategy not only performs well in large time windows,but also performs better in small windows,because it can more accurately predict video content with dramatically changing popularity.Compared with LRU,SW-LFU,JW-LFU and FIFO-LFU cache replacement strategies,this strategy performs better in average hit rate and can effectively improve cache space utilization.