Research On CDN Cache Management System Based On Selective Markov Model

CDN (Content Delivery Network): A system of distributed content on a large intranet or the public internet in which copies of content are replicated and cached throughout the network. When content is replicated throughout the country, or throughout the world, users can have quicker access from the nearest and the best site. But in recent years, with the standard of high definition video and rapid development of broad band, CDN platform is confronted with the problem of how to manage to caching large-scale streaming media.This paper designs a two-level caching architecture of CDN based on combining segment caching and copy caching, through researching on user's access behavior and existing segment cache management technology. In this caching architecture, the whole replica of the video file will be stored in the first level proxy servers, and the segment data of the most popular video programs will be cached and managed in the second level proxy servers, which are deployed on the edge of the network and provides service to the great mass of consumer.Therefore, this paper concentrates on the study of the segment caching management strategy with Selective Markov Models. First of all, we made models of users'VCR actions through analyzing statistics of users'historic access data. After optimizing these models with specific state-pruning strategy, we extract the access patterns which represent the consumer's access habits. Lastly combining with LRU-2 caching replacement algorithm we built FPMM_VLike_LRU₂ caching management strategy for streaming media environment.Finally, our strategy will be certified correct and effective by simulation on the SIMULINK simulation platform, compared with the classic algorithm such as LRU-2, SP-LRU-2, and FP-LRU-2 in accessing delay, byte hit ratio, the space complexity and so on.