Web Service Discovery And Selection Mechanism Based On SP2P

The rapid growth in Internet technology provides transient opportunities forenterprises, so it becomes a tendency for enterprises to utilize information service toimprove their transaction. It is a hot issue and challenge to self-manage and deploy servicebased on internet for optimal allocation of resources, to find best trustworthy partners andto form a dynamic strategic alliance under the open computing environment. It is animportant research topic to combine the technology of semantic Peer-to-Peer andtechnology of Web service to realize the self-management and deployment of enterpriseservice and the deployment and discovery of credible service. The thesis centers on themanagement of Web service in semantic P2P and the selection of Web service based onquality of service and model of service credit. The originality and main work of the thesisincludes the following aspects.(1) In order to solve the bottleneck problem of centralized registration center andimprove inter-enterprise service discovery performance, an unstructured topology serviceindex aggregation is designed with its backup structure and topology network maintenancealgorithm. The thesis finds the semantic mode matching process is a common problem inP2P application to be addressed. The paper uses the technique of the Bloom filter to buildweb service description pattern aggregation process, designs web service pattern discoveryalgorithms under the P2P environment, and actualizes the sameness between the inquiriessemantics and the routing semantics. By experimental comparison and contrast analysis, itachieves fast location of nodes based on semantic P2P service registry.(2) To solve the problem of area spatial information service integration based onsemantic and actualize the service discovery based on semantic, the semantic sharingframework of GIS services based on the P2P environment is designed by sub-ontology andsubspace structure, the semantic integration process and series of needed ontologies aredesigned. The paper proposes the sub-ontology similarity computing algorithm byinformation theory to combine the model of semantic distance similarity in order toconstruct the information services similarity algorithm. The experiment proves the new services similarity algorithm has higher performance than classic one.(3)To solve the problem of matching precision and selection of individual demand ofWeb service based on service quality. The paper puts forward a new three-level QoSontology model to meet the individual demand for service quality and the extendabledemand for property of service quality. The thesis designs a semantic matching algorithmbased on QoS parameter value-code and designed a matching algorithm which is based onQoS attributes parameter values. The effectiveness of the QoS parameter values similarityselection algorithm is verified by experiment, which provides the closest service to meetthe demand of the user in quality of service.(4) To use the service under unreliable network environment and resist customers’venomous behavior, the paper proposes a credit model of P2P based on the reputation. Thethesis designs the service reputation model and the credibility model of service evaluatorwith time window factors to design the appropriate punishment function and theencouraging function to perfect the above two models. Through experimental analysis, itis found that the credibility model supports users to select reliable service and has abilityto resist malicious evaluation.