The Study Of Web Service Discovery Method Based On Graph

Web service is a new bread of Web applications; it is the core of SOA (Service-Oriented Architecture), in which there are three different roles, i.e., Service providers, Service broker and Service requestors, and three basic operations, i.e., Web Service publication, Web Service discovery, Web Service binding/invoking. Service discovery is a key process in the whole procedure. Traditional Web service discovery processes use match algorithms to look up in UDDI registration center according to the functionalities of Web services, and then sort the satisfied services according to their Quality of Service (QoS), finally choose the most suitable one. However, to discover a suitable Web service from the repository is a time-consuming task, since the search space is in general exponentially increasing. Thus, the efficiency of Web service discovery becomes more and more important. In previous literatures, some researchers proposed a service discovery algorithm based on graph search. Based on these results, we are enlightened by the research on Web graph, proposed a new concept and framework for service discovery, i.e., Web Service Graph (hereafter this is abbreviated as WSG) which is a semantic index for Web service over the UDDI repository. The simulated experiments show that the newly proposed framework works well for service discovery. The efficiency and response time of service discovery can be improved greatly without undermining other measurement indicators.This paper is organized as follows:1. It fisrt outlines Web service basic conceptions, its system structure, quality appraisal and so on. Especially, it introduces the basic concepts, basic techniques and some existing questions related with Web service discovery.2. Then, based on the related graph theory and technology, this paper studies the way on how to construct Web service graph in UDDI registration center and the way on how to effectively store it.3. After that, it studies the service discovery algorithm based on Web service graph, and carries on some QoS comparisons between traditional Web service discovery method and WSG-based service discovery method through massive simulations, demonstrates the superiority of WSG-based service discovery method. In addition, we make some improvements on the algorithm according to the actual situation, for example, starting from the biggest degree point to match, starting from the essential point to match etc.4. Based on the study of UDDI registration center and Web service graph, we choose the open source JUDDI platform and expand it to support Web service graph, so a true application platform will be built. In addition, we introduce the knowledge map model, modeling the relationships between Web services; this can solve the problem of how to visualize Web service composition flows and effectively transmit them on network, how to interactively select for consumers to choose their preferences, and so on.