Pervasive Computing Based On Semantic Service Discovery

Pervasive computing, sometimes called ubiquitous computing, is the third wave in computing. Locating a network service or a device on demand is a challenging task for pervasive computing. This paper provides an overview and comparison of several prominent service discovery mechanisms currently available, includes Jini, UPnP, Salutation, SLP and so on. Most of these systems use simple interface-based or attributed-based matching. They lack expressive languages, representations and tools that are good at representing a broad range of service descriptions and are good for reasoning about the functionalities and the capabilities of the services. Common ontology infrastructures are often missing from the existing service discovery architectures. In this paper we introduce a new approach that explores the use of Semantic Web language in building architecture for service discovery in pervasive computing environment. We propose an OWL (Web Ontology Language) based ontology for service description, and for supporting logic-based service matchmaking. By means of OWL we are able to define the "purpose"of a service by giving it a semantic meaning. Additionally, the ontology shows the semantic relationship of the services. We also point out how to describe a service provider and a service requester using the ontology. The ontology system was divided into several parts, includes space, time, personal information and service. Besides service description, Service matchmaking is another important problem in service discovery. We introduced a semantic matching algorithm according to the ontology. The algorithm takes the semantic meaning of service into account, enable interact between user and device or between device and device. It improves the veracity of service matching, and paves the way for fuzzy matching and service composition. The algorithm classifies and compares the request class, inputs and outputs to the advertisements class, inputs and outputs. The algorithm different from the most existing service discovery algorithm because the reasoner can reveal the semantic hierarchy of the classes, inputs and outputs, avoid the simple interface-based or attributed-based matching. At last, we implement the algorithm based on Jena, JESS and OWLJessKB.