Research On Chord Protocol Based Peer-to-Peer Network Model And Searching Technology

P2P (Peer-to-Peer) technology has been seen as one of the most popularcomputer technology in twenty-one century. With the network technologyadvances and power of PCs enhances, the computing pattern of Internet istransforming from C/S to P2P. High efficiency resource search mechanism is theresearch keystone of the P2P technology. The paper reviews P2P's basic conception,network architecture and the primary application domain. The primary searchingalgorithms at present are summarized. The searching algorithms of unstructuredP2P networks and structured P2P networks are analyzed and pointed out theiradvantages and disadvantages.Structured P2P network is the third generation P2P networks and its DHTsearch algorithm has the higher search efficiency which is suitable to large-scaleP2P networks and its expansibility is very good. It is the study hotspot at present.But structured P2P networks Chord have not consider the actual physics addresswhen constructing networks which lead to the difference between coveringnetwork and bottom actual network is very big. It is the detouring problem ofChord network. In this paper a new P2P network model based on Chord protocolhas been proposed by the deep analyzed of Chord network. The new modelseparated the nodes in the networks into different clusters according to their actualphysics address vicinity. Every cluster is a Chord network. Clusters interconnectedcomposed the distribute P2P network. The new model has considered thedifference of nodes capability by leading into the concept of cluster header node.And in order to enhance the probability of search inside of a cluster and decreasethe number of times search across cluster, the replica technology of cluster header node has been proposed. Therefore the new model not only inherited the advantageof Chord network, but also reduced the volume of network traffic, decreased theroute location expense, enhanced the search efficiency.Owing to the new model characteristic in the respect of topological structureand data distribution, a two-stage hybrid search algorithm has been proposed.This algorithm include cluster inside search and clusters search two-stage hybridimplement. Searching based on cluster and respectively adopt different searchalgorithm according to different topological structure. Inside the cluster we adoptDHT search algorithm, and between the clusters we adopt the random walk searchalgorithm which is the improved search algorithm of the flooding algorithm. Thisalgorithm combined the advantage of the DHT search algorithm and random walksearch algorithm. The experimental results show that the algorithm under newmodel is better than DHT search algorithm.