Content Replica Placement Algortthem Research For CDN-P2P

With the development of the Internet, on the one hand, the content on the Internetchanges from static pictures and text to multimedia services, on the other hand, as theincreasing of Internet users, request frequency of Internet content also increases, whichbrings new challenges to content distribution. Because of the two issues mentionedabove, the Content Delivery Network (CDN) and Peer-to-Peer (P2P) have been paidgreat attention. Both CDN and P2P have their own limitations, but they can complementeach other well in the cost of deployment, scalability, et.al. Therefore, in recent years alot of researches have been conducted on content replica placement, redirect request,content update, billing, et.al of hybrid CDN-P2P. And reasonable replica placementtechnology can reduce the backbone network traffic, improve the hit rate of the edgeservers.Firstly, this article compares three integration programs of CDN-P2P, namelycontrol plane integration program, network overlay integration program and devicelevel integration program. According to the situation that a large number of CDN nodeshave been deployed by the recent companies and IDC, this paper determines the hybridarchitecture of CDN-P2P, which fixes a foundation for content replica placement andsimulation.Next, in order to reduce the transmission cost between the edge server and sourceserver and improve the hit ratio of the edge server, this paper conducts research on thecontent replica placement from the angle of content transmission cost and user’s interestdegree.From the perspective of content transmission cost, in order to improve the hit ratioof the edge server, this paper presents two models and two kinds of algorithms to placethe content on the edge server with bigger transmission cost. Firstly, this paper proposesmodel1, and this model balances the transmission cost between P2P users and the costbetween the edge server and the source server. Then, this paper puts forward model2,used to calculate the transmission cost which produced by the P2P users who directlyrequests content from source server. Based on cost model1and model2, this paper puts forward content replica placement which based on weight seleciton algorithm, used topull the content which has bigger transmission cost to the edge server. At the same time,this paper and, renovates this algorithm and makes the content cost on the edge server ismuch bigger than before. Then simulate these models and algorithms.From the user’s interest degree point of view, in order to reduce the transmissioncost between edge server and sorce server, this paper proposes a replica placementbased on the degree of user interest which puts the content with high degree of userinterest on the edge server. This algorithm uses the principal component analysis to getcomprehensive content features, and then compares the content features and thecomprehensive content features in order to get the degree of the user interest, and putsthe content with high degree of user interest on the edge server. The simulation resultsshow that, compares with the existing algorithms, this algorithm can partly improve thehit ratio of the edge server, and reduce the transmission cost between the edge serverand source server.