Research On P2P Video On-demand Streaming System Caching Mechanism

With the increasing of network users, the improving the PC configuration, increased popularity in network television and the acceleration of the pace of life and people’s online video browsing, P2P video on-demand system is more and more popular. Currently,There are many kinds of P2P video on-deman system caching mechanism which are proposed at home and abroad.However,most of these caching mechanism does not consider the information of local nodes or solely focus on the aspect of the local information.While taking into account the local information,they need to use a lot of server overhead.The paper presents a caching mechanism based on user behavior,the mechanism not only takes the local information into account,but also considers the global information. Caching mechanism based on user behavior combined with the general behavior of the user,the personal behavior of the user,program subscriptions and favorites,channel cooperation,pre-download and pre-upload,considering the general behavior of the user,the personal behavior of the user,program subscriptions and favorites and other factors to determine the ratings of video segments.Cache the high score video segments during the pre-downloading,replace the low score segments when the client cache is full.User behavior-based caching mechanism to measure the advantages and disadvantages of the cache policy from the system hit rate,server loading,network loading and locals hits.The author combines the Brite and PeerSim achieving a P2P video on-demand simulation system, implementing the user behavior-based caching mechanism which is proposed in this paper,the probability-based caching mechanism,the popularity-based caching mechanism,first in and first out caching mechanism. The results show that the user behavior-based caching mechanism increases the cache hit rate of8.82%, reduces the server loading of4.05%, nearly raises3times the local hits.