Rule-based context-aware service discovery based on service publicness

Abstract

Service discovery usually adapted context information to rank relevant services. Most proposed service discovery considers very limited kinds of context, such as geographical location and user preference. However, location and preference alone is not enough to filter out inappropriate service in pervasive computing environments. In this thesis, we present a new compound context, publicness and use it to filter out inappropriate services. Publicness is decided by multiple contexts, those that related to place, user, and information. With publicness context, proposed service discovery scheme filter out inappropriate services which cannot be achieved by previous approaches with negligible overhead. In addition, as user context changes, this scheme cannot filter out precisely in some situation. Then, this service discovery scheme performs post processing for accuracy.