Efficient and Flexible Sensornet Checkpointing

Abstract. Developing sensornet software is difficult partly because of the limited visibility of the system state of deployed nodes. Sensor-net checkpointing is a method that allows developers to save and re-store full system state of nodes. We present four extensions to sensornet checkpointing—compression, binary diffs, selective checkpointing, and checkpoint inspection—that reduce the time required for checkpointing operations considerably, and improve the granularity at which system state can be examined and manipulated down to the variable level. We show through an experimental evaluation that the checkpoint sizes can be reduced by 70%-93%, and the time can be reduced by at least 50% because of these improvements. The reduced time and increased gran-ularity benefits multiple checkpointing use cases, including automated testing, network visualization, and software debugging.