This paper presents Pixie, a new sensor node operating system designed to support the needs of data-intensive applications. These applications, which include high-resolution monitoring of acoustic, seismic, acceleration, and other signals, involve high data rates and extensive in-network processing. Given the fundamentally resource-limited nature of sensor networks, a pressing concern for such applications is their ability to receive feedback on, and adapt their behavior to, fluctuations in both resource availability and load.
The Pixie OS is based on a dataflow programming model based on the concept of resource tickets, a core abstraction for representing resource availability and reservations. By giving the system visibility and fine-grained control over resource management, a broad range of policies can be implemented. To shield application programmers from the burden of managing these details, Pixie provides a suite of resource brokers, which mediate between low-level physical resources and higher-level application demands. Pixie is implemented in NesC and supports limited backwards compatibility with TinyOS.
We describe Pixie in the context of two applications: limb motion analysis for patients undergoing treatment for motion disorders, and acoustic target detection using a network of microphones. We present a range of experiments demonstrating Pixie’s ability to accurately account for resource availability at runtime and enable a range of both generic and application-specific adaptations.