Analysis On Discovery Mechanism Of Distributed Web Service Based On Peer-to-peer Alliance

Web resources show its characteristics of cross-platform, dynamicness,large-scale, even uncontrolled nature with its increasing in the network.Peer-to-Peer technology and Web service technology are connected necessarily inorder to discover different resources which are shaped and shared by Web servicein the new network environment.The unstructured Peer-to-Peer networks meet the requirements of thestructure in the distributed Web Service discovery. The issue of the distributedWeb Service discovery is converted to the method of finding the peers whichprovide Web services.A distributed Web service discovery method named W-Search was proposedin this dissertation which is based on Peer-to-Peer alliance. The peers whichprovide Web services aggregated into Peer-to-Peer alliances by their samefunctional properties. Peers published their local Web services and saved remoteservices which came from the nodes in the same alliance.Alliance known list and node known list were designed in the query packetstructure. The query packet remembered known alliances and nodes as heuristicinformation. These known lists determined the next hopping of these packets inorder to prevent the same query. The size of known list was under control by usingthe memory window mechanism. The query packets of the same query exchangeinformation over short distance. Known lists would be merged in bumping.First, the query packets of Web services search the alliance which couldfulfill their requirements of the functional properties during roaming in peers.After that, the QoS properties of the query packets are proliferated through thealliance. The best service is elected out according to the Topology Mismatchproblem and the QoS properties. Eventually, the whole Web service discovery getsrealized.The experimental results showed that, compared with the random walkalgorithm, W-Search algorithm saved the delay time and bandwidth costs in theprobing process, and improved the efficiency of services searching. The memorysize of query packets was under control while maintaining efficiency algorithm.