Research On Web Services Composition Framework Based On Agent Organization

As the representative of next-generation-Web technology, Web services can effectively solve such problems as the complexity of software interface, difficulties in updating and integrating, etc, so the number of Web services on the network is rapid growing. However, single Web service is very difficult to meet the actual needs of user, how to use existing Web services to meet the needs of practical application, it is necessary to compose several Web services together to bring into play the real potential of Web services. At present the research about Web services composition has made great progress, but with the rapid development of Internet, the changes of environment, and the changes of user's needs, Web services composition still faces many challenges, such as the passive of Web services, the centralized processing, the lack of adaptability and the intervention dependence, etc.To deal with the problems discussed above, this paper propose an Agent organization based framework for service composition, by analyzing some key technologies, discussing some related research projects, introducing the software Agent technology, and running support with Agent organization. In the framework of the system: Firstly, to enhance the autonomy and adaptation of service, software Agent represents as Web service. Secondly, to research the interaction and coordination between Agents a method through the Agent organization and Agent conversation is used. Thirdly, context-based Web services selection is used to achieve automated selection. Lastly, Web service composition system has certain self-adaptive by exception handling. This paper is outspreaded as following five aspects:(1)The design of service composition frameworkAbove all, an Agent organization based framework for self-adaptive Web service composition is presented. Its elements as well as the relationship of them are all discussed in detail. Next, context information and its transmission between Agents and Web services are discussed in detail. Next, a context-aware Agent model and its structure and expression are discussed. Lastly, the conversation between Agents is presented.(2)Agent organization An Agent organization based method for Web services composition is proposed. We research the business process of Web services composition and its interaction through Agent organization. And research Agent organization through organization structure and organization norm. Before it an Agent organization definition is presented. And a method of Agent organization formation based the business process of Web services composition is proposed.(3)Context-based Web services automatic selectionTo deal with the confusion of tradition Web services selection, a context-based method is proposed. This method is according to single Web service context and user's context to automatically select the most suitable for user. A QoS mode and trust model in Web services context and their calculation methods is discussed in detail. Then, WSDL is expanded so that it can contain QoS information. The main ideas and process description of automatically selection based single Web service context and user's context is discussed.(4)Exception handlingIn order to improve the adaptability of services composition, exception handling is researched. To achieve adaptive running, exception handling mechanism is discussed, so that the system can go on running when encounters problems.(5)Web services composition platformThe design and implementation of the services composition framework is presented. This dissertation introduces the process of requirement analysis, system design framework, Agent organization, Agent model and implement, and so on. With an instance, the general design, detailed design, and the execution mechanism which is implemented are emphasized.Simulation results indicate that this framework can support highly concurrent requests and large volume of data more effectively than traditional architecture, so it is helpful to improve the scalability of composite services.