Research On Mechanism Of Resources Search And Location In P2p System

As the P2P networks extensive application in people's lives, P2P network management technology become the current hot issues in the P2P research. The mechanism of resource location technology in P2P system is the critical technologies in P2P applications. The mechanism of resource location technology in P2P system related to several aspects of implementation such as the node to join, keyword query, the node's stability and failure management, and the withdrawal of the node.The typical algorithms of the P2P network are analysised in this thesis. The typical algorithms of unstructured P2P networks mainly use flooding or roaming mode, resulting in excessive consumption of network bandwidth or randomness choosing neighbors. Typical structured algorithms of P2P network are DHT-based algorithms, have adopted the SHA-1 hash function to obtain the identifiers, and are routed using the finger table, therefore are more stable network structure and higher the query efficiency than unstructured algorithm. In order to improve query efficiency of chord algorithm that is a typical algorithm of the structured P2P, this paper presents BPT CHORD model, and the model's basic idea is to first set the threshold according to the calculation of node-to-point latency, delineate the corresponding domain packet and named it the O-chord. Within the O-chord use traditional algorithms to build finger table for the nodes. And then set the threshold value, for the first node to join of each group automatically become a super node. All the super-nodes consist S-chord, and every super-node construct a bi-directional finger table . In case of the O-chord fail to find resource, through the super-node relevant flag returned the information of failure to find the resources ,then the super-node decidie which group domain to lookup through the bi-directional index table .Thus ,effectively solve the problem of the chord ring exists that in a logical ID of the two adjacent nodes may not be adjacent in the physical path, as well as one-way query to bring the problem of low efficiency.Simulation results show that the model algorithm to improve the average query path length, the average query delay, the average extension rate of three indicators of improvement, so effectively improve the efficiency of resource location. At the same time can provide a reference for further theoretical research.