A System Framework for Web Service Semantic and Automatic Orchestration

In this paper, we present the framework of semantic and automatic service orchestration (SASO) system for Web services modeling and composition. The SASO system has the following features: 1) it adopts a semantic approach to model Web services, and 2) it supports a Web service selection policy based on user preferences. First, we express services in their semantics so that relations among the services can be identified through ontology modeling, which is to be accomplished in two steps: input/output parameter modeling and Web service modeling. Then, we provide an algorithm for automatic service composition with a selection policy that addresses user priority. This relieves users from the tasks of manually browsing, monitoring, and composing Web services. Furthermore, SASO supports fuzzy composition of Web services when exact Web services cannot be found. In SASO, composition plans are represented by graphs that can be interactively modified. These graphs are compiled into executable code to carry out the composite service. Our empirical results demonstrate that SASO not only finds the appropriate services for composition but also the best plan to execute them.