I. Introduction
As computing and communication technologies get more sophisticated, intelligent personalized services are possible by rapidly reflecting the context of the user's environment in the service. In the development of context-aware systems, various factors need to be clearly specified, including the context and the relationship among them, user preference, sequence of events, response time, etc. The development process becomes even challenging if high quality of service needs to be guaranteed, particularly for time-critical applications [1], [4]. A model simplifying the manipulation of complex context-aware service is highly desired for the specification and analysis of the service.