Research On Non - Structured P2P Network Resource Search Technology Based On Node Interest

In recent years, the research of P2P network (Peer to Peer network) technology is becoming a hot topic as the rapid development of Internet users and Internet technology and the rapid expansion of resource information. The different between P2P network and the traditional C/S(Client/Server) centralized network model is better scalability, self-organization,fault tolerance ability and other advantages,and there is no longer only a central server.All the peers in the network are the same roles,not only can act as a client role but also can act as a server role,which can provide resources for other peers while get the necessary resources from other peers.However,with the number of Internet users rapid growing,how to find the necessary resources quickly with low spent in the massive resources is the main difficult problem.The paper works in the topology and resource search algorithms of unstructured P2P networks, as following three areas:1.On the basis of the analyze the existing research algorithms,and the peers of unstructured P2P network are:dynamic,loosely,no fixed logic address,this paper proposes a kind of complete binary tree topology of unstructured P2P networks which bases on the peer of interest.As every peer in the P2P network has its interesting things than establish a logical connection for the peers which has high interest similarity value, and choice the best comprehensive performance super node as the root node of a complete binary tree,combine with the architectural feature of complete binary tree establish a kind of complete binary tree topology of unstructured P2P networks which bases on the peer of interest,the higher interest similarity value of peers the more near in the tree.In the processing of search resource,can priority forward and query the higher interest similarity values can effectively reduce resource spent in the irrelevant peers.2.Based on the last chapter proposes a kind of resource search algorithm is interval type bidirectional search algorithm,divide the peers into the normal peers and the query peers,the query peers store the resource index information of his child peers and the address information of his grandchildren peers.In the resource search processing can skip the child peers but forward the query messages to the grandchildren peers,can optimize the total amount of messages generated because of forward too many times to some certain extent.3.At last using PeerSim simulator conducted experiments and simulations the proposed algorithms,by Comparing the total amount of messages generated,search success rate,the average path length and others to verify the effectiveness.