Research On Semantic Service Composition And Its Supporting System

Web Services as an emerging model of applications and remote access standards is becoming a good solution to the application and integration of data across different platforms. It can integrate software that developed by deferent organizations into services by using open standards such as WSDL, UDDI and SOAP. More importantly, it could organize the existent services to achieve more complex business functions, so it has become the focus of current research and hotspot on how to compose Web Services to provide more valuable "composite service". WS-BPEL 2.0 is an XML-based language and has become the actual standard of Service composition language. It has a high interoperability, cross-platform and the characteristics of loosely coupled. As its services description capability is limited in the language level, so it could not express the semantic information and doesn't provide any means to monitor the QoS of the Web Services.In this paper, how to do the BPEL composed services is analyzed in detail. The adaptation of BPEL to composite Web Services lacks dynamicity, semantics, and monitoring of the QoS. In order to solve these problems, we introduce the OWL-S and extend it with QoS attribute to describe the attribute and functions of Web Services. An ontology-based three-stage matching algorithm is proposed as well. The algorithm can match service by using the service profile, QoS, and input and output of the service .It greatly enhances the speed of matching service under the premise of ensuring precision; According to the fact that the requested service is not easy to find the appropriate service, a recursion-based reverse search chain algorithm is proposed. The algorithm can be used to transform semantic, do reasoning in the existing services and construct the satisfactory service chain when the expected service can not be found.Apache ODE is an open-source workflow engine; it can hold the industry-standard WS-BPEL 2.0 compatible with the old standard BPEL4WS 1.1. It can only passively implement the business process described by the BPEL and can not find the service initiatively .It also lacks the support of semantic. OWL-S can describe the semantic of the Web Services and can do service reasoning, composition and execution automatically. In this paper, the semantic matching proxy for Apache ODE is extended. The designed system can do Service composition dynamically and semantically. Finally, a prototype system is implemented and up run in the Lab to prove that the work of this paper is feasible and effective.