The use of Web services revolutionizes the way that many e-commerce, consumer software, scientific computing, and telecommunications applications are provided. The key to realizing the full potential of the Web services paradigm will be automated discovery, composition, analysis, and monitoring of web services. Achieving this requires advances in many aspects of computer science, including programming languages, data management, workflow, transaction management, distributed computing, knowledge representation and automated reasoning. This project provides key foundations and advances in data and information management needed for success of the Web services paradigm. This project expands on existing work to develop models of the semantic and behavioral characteristics of web services. The automata-based models can specify a "behavioral signature", that represents abstractions of both the "local" behavior of individual Web services and the "global" behavior of Web service compositions. Using semantics, behavior and I/O signatures, the project develops new approaches to automated composition, and studies global properties of compositions, including characterizations of the global behavior and static analysis of how data is passed and manipulated. It also lays the foundations for the query languages and frameworks useful in supporting automated web service discovery and composition in a scalable context. In addition to the study of formal models, the research develops various software tools that will be available via the project's Web site (www.cs.ucsb.edu/~su/NSF0415195/). The resulting formalism will improve the quality of deployed web services since properties of their behavior can be verified, and foster web service adoption and reuse since compositions of services could be synthesized and analyzed. This is a collaborative university (Jianwen Su, University of California, Santa Barbara)/industry (Rick Hull, Lucent) research project. This will facilitate transferring the theoretical results to practice and finding realistic applications to test the techniques. The project exposes students and academic researchers to industry needs and industry practitioners to university research through a plan of mutual visits.
