Design principles for resource management systems for intelligent spaces.pdfVIP

To appear in the Proceedings of The Second International Workshop on Self-Adaptive Software (IWSAS’01) Budapest, Hungary, 2001 Design Principles For Resource Management Systems For Intelligent Spaces Krzysztof Gajos, Luke Weisman, and Howard Shrobe MIT AI Lab, Cambridge, MA, USA, , Abstract. The idea of ubiquitous computing and smart environments is no longer a dream and has long become a serious area of research and soon this technology will start entering our every day lives. There are two major obstacles that prevent this technology from spreading. First, different smart spaces are equipped with very different kinds of devices (e.g. a projector vs. a computer monitor, vs. a TV set). Second, multiple applications running in a space at the same time inevitably contend for those devices and other scarce resources. The underlying software in a smart space needs to provide tools for self-adaptivity in that it shields the rest of the software from the physical constraints of the space, and that it dynamically adjusts the allocation of scarce resources as the number and priorities of active tasks change. We argue that a resource manager can provide the necessary functional- ity. This paper presents a set of guiding principles for building high-level resource management tools for smart spaces. We present conclusions we arrived at after two years of exploring the topic in the Intelligent Room Project at the MIT AI Lab. The paper is based on a number of imple- mented and tested tools. 1 Introduction For several years, our research group in the MIT AI Lab has been developing an “Intelligent Room” [8, 9, 5], a space that interacts wit