| Wireless sensor network (WSN), a system of networked miniature devices with sensing, processing and storage capabilities, has emerged as a promising technology to instrument our environment, monitoring and actuating a large variety of environmental phenomena. Wireless sensor networks have found their usage in a large spectrum of applications, including habitat monitoring, urban sensing and disaster control.; However, despite the potential for transformative scientific and even social change that wireless sensor networks seem to promise, their development is, at present, still nascent. The major obstacles are their resource constraints, dynamism and unreliability of working condition and high labor cost for deployment. To overcome these difficulties, numerous simulation tools have been built, given simulation as a cost-effective, controllable and flexible development tool. Unfortunately, none of them is satisfactory in terms of fidelity, scalability and debugging capability.; In this dissertation, we discuss our work in an attempt to build an easy for debugging development system based on scalable, high fidelity simulation. The key component of our system is a scalable, distributed full-system simulator for a variety of sensor devices, including motes and Stargate. Our performance study shows that the simulator can scale to thousands of sensor nodes with cycle accuracy. Built upon that, we design and implement debugging services at different levels of a sensor network, including debugging points and virtual debugging hardware for individual sensor device, coordinated break for multiple nodes and checkpointing service for a complete network. We also extend our system to support a seamless fusion of simulation and physical networks, a simulation-based augmented reality, which enables novel development methodologies. |
