Distributed management of contextual affinities in context-aware systems

Abstract

In context-aware computing, distributed information about entities, such as people, places and objects, is captured and made available to applications, which utilize this context. With contexts from multiple entities available, the degree of alikeness of those contexts poses an interesting piece of information. We have tackled the issue of contextual proximity by employing mechanisms known from location-based services on conventional context-aware computing. This paper discusses the idea of the Contextual Map, a context model inspired by conventional map-models in location-aware computing. The Contextual Map represents contextual information in a multidimensional map model, where proximate representations of contexts imply affinity between those contexts. We consider entities, which compose such contexts, are potentially distributed and mobile. This highly dynamic environment puts our focus on the issue how those context carriers are distributed and which architectural conclusions are to be drawn.