Automated Service Brokering in Serive-oriented Architectures (ADDO)

(Funded by Deutsche Forschungsgemeinschaft)

Modern service-oriented architectures support the integration of externally provided services as components in local applications. This integration is primarily achieved using Web Services technology based on standardised Internet protocols. A fundamental challenge is the discovery and linking of appropriate services. This task was solved already in service trading extensions of CORBA and other platforms. However, new developments such as Semantic Web, Quality of Service guarantees, and Web Services have changed fundamentally the technical context and potential of service matching. They promise substantial improvements, but also pose a number of interesting and challenging research questions.

The ADDO project builds on existing results in service trading and exploits new technological achievements for the development of enhanced service trading solutions. Key aspects of the research activities are:

• Using the new service description language OWL-S and based on semantical information provided in service ontologies a service specification technique will be developed that includes quality of service requirements.
• A flexibly customizable service matching algorithm will be developed for the matching of (potentially incomplete and fuzzy) service specifications.
• The trading of service aggregates will be automated. Service composition will be handled transparently for the requesting clients.
• The viability of the results will be evaluated using a prototype implementation of the broker in a realistic application scenario.

Automated service trading and service binding are important building blocks for the self-organisation and self-configuration in distributed systems.