Abstract not found

Many proposed applications for Cyber-Physical Systems will be implemented on a network of computing devices, communicating wirelessly using local broadcast. The network may be ad hoc, in that the participating nodes are not known a priori and must self-configure into a coherent network, and dynamic, in that the participating nodes can change over time as nodes enter and leave the area of interest, fail, and recover. Currently, it is difficult for programmers to write, debug, and verify applications that will run robustly under such conditions. These practioners would benefit significantly from clean, well-defined abstraction layers that can mask some of the failures, uncertainty, and change inherent to this setting. It is up to the research community to produce these abstraction layers. In fact, we believe that the development of good abstraction layers for dynamic ad hoc networks should be a major goal of any Cyber-Physical Systems research initiative. The mobile network research community is currently producing communication abstractions such as IP-style point-to-point routing for mobile ad hoc networks, and reliable broadcast dissemination. Communication abstractions alone, however, are not enough for systems that need to interact meaningfully with a real-world environment. Consider, for example, a first responder network that needs to monitor temperature changes over time within a burning building, or a campus-based PDA network that needs to maintain “virtual graffiti walls ” at several well-trafficked locations. In both scenarios, reliable communication alone does not greatly simplify the task. To facilitate the development of Cyber-Physical Systems we need abstractions that encapsulate data-management functions, such as data collection and dissemination, data storage, and query processing. We also need abstractions that encapsulate planning and control, and an overall architecture to fit these abstractions together into a coherent whole.