Research On P2P Networks Balance Load Optimization Algorithm-Based On File Classification And Implementation

The introduction of the network technology brings about the leapfrog development of e-commerce, which incorporates a large number of commercial resources. For better sharing resource, P2P technology taking resource-sharing as purpose has been applied to e-commerce gradually. However, the effective utility of date would be impossible without a competent data search mechanism. Therefore, the research on data searching in Peer-to-Peer network has important practical and disquisitive meaning. First, this text proposed a new calculation model, applying P2P technology to e-commerce to form a distributed e-commerce registration service system. The distributed registration service system is based on Chord. But as Chord can't be applied to e-commerce directly, more route storehouse should be added into each route form to extend Chord route algorithm. Second, the thesis proposed to use XML file as system metadata to be the format of e-commerce unit releasing and network resources inquiring. In order to balance the commercial resource release load at the nodes of system and to prevent inquiry congestion, the system should adopts the load balancing tactics of two choices to realize the services of a high-efficient distribution registration center of a load balancing. Third, the text put forward an improved algorithm FTChord(FT-File Type) based on Chord algorithm by analyzing several existing classic algorithm in the Peer to Peer network.The essence of FTChord is to classify files according to their type so that the release of and search for a file can be conducted according to the type of the file. FTChord algorithm adopts a new method of hash Nhash (N-New ), to replace the existing Hash method. Files of the same type can be Nhashed onto some specific key value, hence the hunting zone is narrowed down, and the efficiency and accuracy of searched is improved. Experiment proves that compare with the initial simulated node scheme of Chord, using two choices algorithm torealise system load balance needs less shared route information stored on the nodes. The result of comparism experiment also indicates that the search efficiency based on FTChord algorithm improves a lot compared with that based on the original Chord protocol.