Architecture and Performance of a Scalable Telemetry Acquisition and Distribution Infrastructure

As the cost of embedded devices, sensors, and wireless networking decreases, the economic viability of intel- ligent services that sense conditions in the physical world and trigger responses to them becomes increas- ingly attractive. Sensor networks for environmental monitoring and an urban area's population of trucks and automobiles are examples of existing physical infrastructures that can be supported by such services. We have developed a services-oriented, specification- based, ubiquitous computing platform called TOPAZ that abstracts common ubiquitous computing functions and makes them accessible to any application provider through Web-service-based, metered interfaces. In this paper, we describe a scalable architecture of teleme- try service that acquires, processes, and distributes telemetry data and analyze its performance results. Our results demonstrate that through intelligent ag- gregation of event streams and controlled object allo- cation schemes, it is possible to support high event rates using off-the-shelf Web application-server soft- ware and hardware.