A formal consensus-based distributed monitoring approach for mobile IoT networks

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.International audienceInternet of Things (IoT) represent a significant area of network research due to the many opportunities derived from the problematics and applications. The most recurring problematics are the mobility, the availability and also the limited resources. A well-known interest in networks and therefore in IoT is to monitor properties of the network and nodes [1, 2]. The problematics can have a significant impact on the monitoring efforts. Mobility and availability can create incomplete results for the monitoring. It can also represent a challenge to monitor distributed properties. The literature states that accuracy is not always reliable and difficult to achieve due to dynamic properties of the IoT in particular with M2M communications and mobile devices. Therefore we propose a distributed monitoring architecture that relies on multiple points of observation. It provides a consensus mechanism that allows it to aggregate and provides a more meaningful and accurate result. We support our proposal with numerous mathematical definitions that model local results for a single node and global results for the network. Finally, we evaluate our architecture with an emulator that relies on AWS, NS3, and Docker with varying number of nodes, network size, network density, speed, mobility algorithms and timeouts. We obtain very promising results, especially regarding accuracy