Models and Infrastructures for Pervasive Computing

The eld of Pervasive Computing has emerged as an active research area in the past few years [8]. Pervasive computing refers to computing environments which contain millions of computing devices or information appliances [7], embodied in the environment we live in (e.g., our cars, watches and clothes, mobile phones, microwave ovens, refrigerators. etc.). These devices interact with us using our own languages and among themselves transparently and adaptively to provide the services we need often without explicit requests from us. Pervasive computing represents a vision and grand challenges integrating research in areas from user interface, web and database technologies, and all the way down to hardware and communication network. The applications of pervasive computing are only limited by our imagination. We can think of refrigerators that can place an order to the supermarket with the best price when an item runs out. Likewise, terrible stories could be told on privacy issues, human beings being out-smarted by computers, etc. However, we believe there is a large area in between these extremes that even the most conservative would feel comfortable with. For example, systems that automatically display the pollution index on our watches are important for health care; systems that can automatically deliver the right kind of information to us, depending on when and where we are, will boost our productivity. In this paper, we focus on the technology issues instead of dealing the privacy and social issues. Since pervasive computing is massive in terms of the number of devices and services involved and mobile because devices are likely attached to moving physical objects, we place special attention on the temporal and spatial aspects of modeling, dynamic and complex composition of services, and mechanisms that match requests to services in a scalable manner. We point out two important aspects of pervasive computing, namely, model and infrastructure. A logical model for computing devices and services captures features that are essential for devices to connect to services and the recursive composition of services. A basic requirement of pervasive computing is that these devices may exhibit di erent behaviors depending on the location and time. As such, the notion of temporal and spatial properties, including their inter-dependence and impacts on the services provided by the system needs further study. The infrastructure supports devices transparent access to services. This includes the notion of data-centric routing and service community to allow the pervasive computing environment to scale up to potentially billions of devices.