A generalized distributed consensus algorithm for monitoring and decision making in the IoT

In this paper, we propose a method to distributively monitor a dynamic mobile network. For this purpose, we take advantage of the consensus theory to provide each node with a common view of the network. More specifically, we give a decentralized algorithm to estimate a time-varying distribution, where each node has a partial information on this distribution. Our algorithm allows a trade-off between the precision of its estimation and its bandwidth consumption. We validate our approach by simulation under NS3, considering the distribution of several network metrics. Simulation results demonstrate how our algorithm can give insights on the network behavior that could be exploitable in a decision making process.