Energy-efficient monitoring of distributed system resources for self-organizing sensor networks

We study the need for integrated monitoring of system resources for supporting optimizations in self-organizing sensor networks. We focus on energy constrained sensor networks that, in order to improve the network lifetime, may need to reconfigure parts of the network. In general, the quality of the reconfiguration process strongly depends on the accuracy and freshness of information about available system resources such as the remaining energy of the nodes. Classical approaches to network management require a parallel monitoring infrastructure or at least continuous transfer of state information. Piggybacking of monitoring seems to be candidate solution, yet, it turned out that timeliness and overhead represent critical challenges. We present an adaptive and integrated monitoring solution that, substantiated by our simulation results, provide means for efficient transfer of monitoring data piggybacked to application or other monitoring packets with predefined upper latency bounds.