In this paper we present a novel, ontology-based approach to service discovery, which exploits domain knowledge and semantic service descriptions to guide the service discovery process and provide advice on service selection and instantiation in interoperable adaptive information systems. The proposed system architecture for service discovery and advice has the advantage of providing specific advice at multiple levels of granularity during the service composition process. At the highest level, the system can help to determine what kind of abstract service is required against a contextual functional request. Once all the services that can fulfill the required function are discovered, the advice system can recommend an appropriate concrete service, taking into account both problem characteristics and quality considerations. More specialized, in-depth advice can also be given, for example, on how to initialize and configure the parameters of a service. The approach and prototype have been proposed to demonstrate practical benefits in the framework of the MAIS (Multi-channel Adaptive Information Systems) project [9].
