Benchmark

1Towards a Pervasive Computing BenchmarkAnand Ranganathan, Jalal Al-Muhtadi, Jacob Biehl, Brian Ziebart, Roy H. Campbell, Brian Bailey University of Illinois at Urbana-Champaign {ranganat, almuhtad, jtbiehl, bziebart, rhc, bpbailey}@uiuc.eduin its infancy, there is no widely accepted notion of the Abstracttask sets that would form a benchmark. Moreover, differ-Pervasive computing allows the coupling of the physi-ent pervasive computing environments have been built for cal world to the information world, and provides a wealth supporting different kinds of tasks. Hence, it becomes of ubiquitous services and applications that allow users, difficult to compare them. To overcome this problem, we machines, data, applications, and physical spaces to in-identify four broad classes of task, which many environ-teract seamlessly with one another. In this paper, we pro-ments would support one or more of these tasks. We try to pose a benchmark for evaluating pervasive computing measure how well a pervasive computing environment environments. These proposed metrics facilitate the as-supports these tasks in terms of usability, system and pro-sessment and evaluation of different aspects of pervasive grammability metrics. The broad classes of tasks that computing and its support for a wide variety of tasks. drive our benchmark are: (1) Presentation tasks: involving 1. Introduction displaying and navigating information (such as slide-Pervasive computing is a new research ...