Run-time fault diagnosis in wireless sensor systems

Software failures in wireless sensor systems are notoriously difficult to debug. Resource constraints in wireless deployments substantially restrict visibility into the root causes of system and application level faults. At the same time, the high deployment cost of wireless sensor systems often far exceeds the cumulative cost of all other sensor hardware, such that software failures that completely disable a node are prohibitively expensive to repair in real-world applications, e.g. by on-site visits to replace or reset nodes. This thesis describes NodeMD, a fault management system designed to improve node debugging capabilities prior to deployment, and enable remote debugging on in-situ sensor nodes that fail. This system successfully implements lightweight run-time detection, logging, and notification of software faults on wireless mote-class devices. NodeMD introduces a debug mode that catches a failure before it completely disables a node and drops the node into a state that enables further diagnosis and correction, thus avoiding on-site redeployment. We present a detailed analysis of NodeMD on real world applications of wireless sensor systems