Supporting generalized context interactions

Abstract

In context-aware computing, applications' behavior is driven by a continually-changing environment. Mobile computing poses unique challenges to context-sensitive applications and middleware, including the ability to run on resource-poor devices like PDAs and the necessity to limit assumptions about the network. Though middlewares exist to provide context-awareness to applications, they do not address the limitations inherent in dynamic mobile environments. This paper discusses a lightweight approach to context-sensitivity that takes into account these considerations. We explore the use of modularization to tailor service discovery policies for applications, as well as leveraging existing language constructs to simplify creation and aggregation of different context types. We also discuss an implementation of these concepts, along with three sample applications that can automatically propagate changes in context to clients running on devices from mobile phones to desktop computers. © Springer-Verlag Berlin Heidelberg 2005.