Roy Shea

A dynamic operating system for sensor nodes

Sensor network nodes exhibit characteristics of both embedded systems and general-purpose systems. They must use little energy and be robust to environmental conditions, while also providing common services that make it easy to write applications. In TinyOS, the current state of the art in sensor node operating systems, reusable components implement common services, but each node runs a single statically-linked system image, making it hard to run multiple applications or incrementally update applications. We present SOS, a new operating system for mote-class sensor nodes that takes a more dynamic point on the design spectrum. SOS consists of dynamically-loaded modules and a common kernel, which implements messaging, dynamic memory, and module loading and unloading, among other services. Modules are not processes: they are scheduled cooperatively and there is no memory protection. Nevertheless, the system protects against common module bugs using techniques such as typed entry points, watchdog timers, and primitive resource garbage collection. Individual modules can be added and removed with minimal system interruption. We describe SOSs design and implementation, discuss tradeoffs, and compare it with TinyOS and with the Maté virtual machine. Our evaluation shows that despite the dynamic nature of SOS and its higher-level kernel interface, its long term total usage nearly identical to that of systems such as Matè and TinyOS.

SensorWare: Programming sensor networks beyond code update and querying

Wireless ad hoc sensor networks have been largely designed with static and custom architectures for specific tasks, thus providing inflexible operation and interaction capabilities. Efforts to make sensor networks dynamically programmable stumble upon the problems of algorithmic expressiveness, compactness of transferred code, efficiency of executed code, and ease of programming. In short, the problem is the choice of abstraction for the sensor node run-time environment. Our framework, called SensorWare, defines and supports lightweight and mobile control scripts that allow the computation, communication, and sensing resources at the sensor nodes to be efficiently harnessed in an application-specific fashion, through the use of abstraction services. A key feature is that the run-time abstraction can change by dynamically defining new services. Furthermore, by making the scripts autonomously mobile we enable the deployment of the algorithm to be tied to its execution, a feature that reduces the code transferred, compared to conventional code deployment and update approaches. The implementation of SensorWare on an XScale-based prototype sensor node platform occupies less than 240 KB of code memory. The implementation is used to measure the delay and memory overheads, but more importantly, quantitatively highlight the trade-offs involved in run-time abstraction versatility.

Temperature Compensated Time Synchronization

Time synchronization in embedded sensor networks is an important service for correlating data between nodes and communication scheduling. While many different approaches to the problem are possible, one major effect of clock frequency difference between nodes, environmental temperature changes, has often been left out of the solution. The common assumption that the temperature is static over a certain period of time is often used as an excuse to assume constant frequency errors in a clock. This assumption forces synchronization protocols to resynchronize too often. While there exists hardware solutions to this problem, their prohibitive high cost and power consumption make them unsuitable for some applications, such as wireless sensor networks. Temperature compensated time synchronization (TCTS) exploits the on-board temperature sensor existing in many sensor network platforms. It uses this temperature sensor to autonomously calibrate the local oscillator and removes effects of environmental temperature changes. This allows a time synchronization protocol to increase its resynchronization period, without loosing synchronization accuracy, and thus saves energy and communication overhead. In addition, TCTS provides a stable clock source when radio communication is impaired. We present the theory behind TCTS, and provide initial results of a simulated comparison of TCTS and the flooding time synchronization protocol.

On the interaction of clocks, power, and synchronization in duty-cycled embedded sensor nodes

The efficiency of the time synchronization service in wireless sensor networks is tightly connected to the design of the radio, the quality of the clocking hardware, and the synchronization algorithm employed. While improvements can be made on all levels of the system, over the last few years most work has focused on the algorithmic level to minimize message exchange and in radio architectures to provide accurate time-stamping mechanisms. Surprisingly, the influences of the underlying clock system and its impact on the overall synchronization accuracy has largely been unstudied. In this work, we investigate the impact of the clocking subsystem on the time synchronization service and address, in particular, the influence of changes in environmental temperature on clock drift in highly duty-cycled wireless sensor nodes. We also develop formulas that help the system architect choose the optimal resynchronization period to achieve a given synchronization accuracy. We find that the synchronization accuracy has a two region behavior. In the first region, the synchronization accuracy is limited by quantization error, while int he second region changes in environmental temperature impact the achievable accuracy. We verify our analytic results in simulation and real hardware experiments.

Scoped identifiers for efficient bit aligned logging

Detailed diagnostic data is a prerequisite for debugging problems and understanding runtime performance in distributed wireless embedded systems. Severe bandwidth limitations, tight timing constraints, and limited program text space hinder the application of standard diagnostic tools within this domain. This work introduces the Log Instrumentation Specification (LIS), which provides a high level logging interface to developers and is able to create extremely compact diagnostic logs. LIS uses a token scoping technique to aggressively compact identifiers that are packed into bit aligned log buffers. LIS is evaluated in the context of recording call traces within a network of wireless sensor nodes. Our evaluation shows that logs generated using LIS require less than 50% of the bandwidth utilized by alternate logging mechanisms. Through microbench-marking of a complete LIS implementation for the TinyOS operating system, we demonstrate that LIS can comfortably fit onto low-end embedded systems. By significantly reducing log bandwidth, LIS enables extraction of a more complete picture of runtime behavior from distributed wireless embedded systems.

Optimizing Bandwidth of Call Traces for Wireless Embedded Systems

Call traces expose runtime behaviors that greatly aid system developers in profiling performance and diagnosing problems within wireless embedded applications. Strict resource constraints limit the volume of trace data that can be handled on embedded devices, especially bandwidth limited wireless embedded systems. We propose two new call trace gathering techniques, local identifier logging and control flow logging, which provide significant reductions in bandwidth consumption compared to the current standard practice of global identifier logging. Intuition into the savings made possible by the proposed trace gathering techniques is provided by an analytical comparison of the bandwidth required by various call tracing approaches. Confirmation of this intuition is demonstrated through experimentation that reveals log bandwidth savings of approximately 85% compared to global identifier logging using flat name spaces, and 35% compared to global identifier logging using optimal Huffman coding.

Movement Analysis in Rock-Climbers

The goal of this demonstration is to present a system developed to calculate the energy generated in the limbs of a rock climber. We are interested in showing the difference between the movements of a beginner, intermediate, and an advanced climber.

Application-specific trace compression for low bandwidth trace logging

This poster introduces an application-specific trace log compression mechanism targeted for execution on wireless sensor network nodes. Trace logs capture sequences of significant events executed on a node to provide visibility into the system. The application-specific compression mechanism exploits static program control flow knowledge to automate insertion of trace statements that capture trace data in a concise form. Initial evaluation reveals that these compressed trace logs, when generated, consume just over a fifth of the space required by standard trace logging techniques.