Research And Implementation Of Recommendations Technology Based On Edge Server And Neighbors Resources In CDN-P2P System

CDN and P2P hybrid system can make use of their respective advantages,at the same time also can to improve their weak place,improve the efficiency of file distribution.In this paper,in order to further the efficiency of the hybrid system,the application of personalized recommendation technology,firstly we propose neighbor node dynamic selection algorithm of the CDN-P2P system,and then we discuss the problem of replica placement in the surrogate server,at last we study the recommended strategies of the system.(1)We propose neighbor node dynamic selection algorithm of the CDN-P2P system fused content recommendation technology,divide nodes into different interests subnet using the idea of the content recommendation,neighbor node dynamically adjust for the dynamic changes of node's interest,improve self-service capabilities between nodes.(2)For limitations of the replica placement in surrogate server,we propose dynamic replica placement algorithm in surrogate server based recommended thinking,algorithm on the basis of the above research and analysis of the subnet interest vector,predict contents that the subnet may be interest and place them on the surrogate server ahead of-time,the algorithm can effectively predict the content of each user needs,improve the system efficiency.(3)In order to effectively utilize the characteristics of P2P networks,to further expand the resources sharing rate in the CDN-P2P system,added the personalized recommendation technology in P2P networks,aiming at the shortcomings of the traditional collaborative filtering recommendation algorithm,we propose a personalized recommendation algorithm adapted to the system characteristics,experiments show that the proposed algorithm can effectively improve the quality of recommendation,it is conducive to play to the characteristics of P2P network,improve resource sharing rate accuracy.