Proximity Based Semantic Service Discovery for Pervasive Business

Ubiquitous information systems (UBIS) adapt current Information System thinking to explicitly differentiate technology between hardware devices and software components in relation to people and process. A dynamic mix of users, mobile devices and software services requires research into new approaches to service discovery (including that within a business domain). Central registries of services and brittle fixed connections to single application services do not provide the necessary mobility and flexibility for an adaptive mobile enterprise. A future Web containing a heterogeneous mix of software services and embedded, mobile devices offers many opportunities to more effectively interact with enterprise applications (extracting information for rendering on appropriate devices at optimal times in appropriate places). This is all premised on the ability to discover appropriate service in a natural way. Service discovery can be found in a number of guises: Hardware oriented discovery, Enterprise Service discovery and Proximity based discovery to name but three. This paper explores how enterprise and proximity based services can come together and be discovered using semantic models and a SPARQL based web service (both within the proximity of the mobile clients). More practically, novel techniques for service selection using proximity based ontology filtering are proposed. The ontology is derived from a number of enterprise applications and then used by the semantic search service to identify appropriate business services. Proximity characteristics are proposed as an optimised means to select the enterprise service and filter the ontology search space. The ontology engineering process explores how proximity can be integrated alongside domain concepts. In order to experiment with service discovery a number of banking applications are used as a test bed. Functionality available within the applications is used as example application services – described using OWL ontology and discovered using SPARQL based semantic search algorithms. The proof of concept discovery platform is presented along with performance metrics relating to a set of investment banking trading and risk services.