Application development powered by rapid, on-demand service composition

Context aware application design models are often based on embedding contextual dependencies - in the form of if-then rules - to specify how applications should react to context changes. In this paper we argue that such approaches are impractical and increase the complexity of context aware applications, due to the large variety of possibly even unanticipated context types and their values. We present the design and implementation of a framework for building context aware applications on-demand, as dynamically composed sequences of calls to services. Our system employs goal-oriented inferencing for assembling composite services, monitors their execution, adapts applications to deal with contextual changes, and enables failure recovery mechanisms that minimise application disruption. By means of experimental evaluation in a realistic infotainment application, we demonstrate that the framework provides an effective, efficient, and scalable solution.