CASD: A Framework for Context Aware Service Discovery and Selection

Abstract

We present the architecture for a framework for semantic web-based service discovery, suitable for integration with relational data-base systems. Existing discovery algorithms often lead to poor results due to limitations of the service description used and lack of domain data. Our framework incorporates context aware service discovery vis a vis dynamic (run-time) update using relational (domain) data from a legacy system. A template for service ontologies is provided, so the user may represent their services and the interrelationships between them. The domain data is represented as an OWL-ontology. The systems takes information in the service and domain ontologies and performs rule based reasoning using rules articulating inter-service dependencies as well as dependencies between services and domain data. An aggregation over service quality properties allows the aggregated selection of the best-suited services. The framework is implemented as a web application following the Service Oriented Architecture; extensive testing shows that the system is robust. Features of the framework are illustrated using a detailed case study for the health-care domain.