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 fulfil 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

Recently, Information Systems evolved towards distributed architectures relying on P2P technologies, where peers share data and functionalities as services to enable interoperability, without a-priori knowledge about each other. In this context, automatic service discovery is a crucial task. We propose Semantic Driven Service Discovery in a P2P scenario, P2P-SDSD. In our framework, peers that use and supply Web Services are equipped with their own ontologies, to maintain semantic descriptions of local services (local knowledge) and links towards peers, that provide semantically related services (network knowledge). Local and network knowledge constitute a semantic service overlay over the P2P network. Semantic-driven service matchmaking techniques are exploited both to organise services in the semantic overlay and to guide service discovery through the P2P network.