Sensorlab2: A monitoring framework for IoT networks

Efficiently monitoring a network requires full observability of each node. The main challenge lies in correlating the monitoring logs coming from all the nodes and reconstructing useful network-wide knowledge such as routing topology or end-to-end performance. We propose a monitoring framework in which nodes report events that are significant to the life of the network, following clear specified semantics. Clear semantics allows maintaining a network model, run on a computer on the side of the network, mirroring the state of the real network. The successive states of this model are archived in a database and can be browsed by the investigation tools instead of querying the real physical network. This is especially beneficial in the case of low bandwidth networks or resource-constrained devices encountered in the Internet of Things. With two different examples, LoRaWAN and OpenWSN, we illustrate how this monitoring framework allows using the same generic tool suite to debug, monitor and tweak a diversity of networks.