A distributed data management scheme for industrial IoT environments

Industrial IoT networks are typically used for monitoring systems and supporting control loops, as well as for movement detection systems, process control and factory automation. To this end, data generated by monitoring IoT devices are collected, elaborated and sent to controllers and actuators. The routing of data from IoT sensors to actuators is an integral part of any large-scale industrial network for maintaining critical delay requirements. Centralised schemes are typically used, whereby data are transferred to a central network controller, from where they are accessed by any other node requiring them. This may result in significant overheads and suboptimal resource consumption. In this paper, we propose a distributed, cooperative Data Management Layer (DML), whereby nodes cooperate to store data within the network. The DML is decoupled yet interacts with the underlying Network Plane. Specifically, given a set of data, the sets of nodes generating and requesting them, and a maximum access delay that requesting nodes can tolerate, the DML efficiently identifies a limited set of proxies in the network where data are stored. Given the mentioned constraints, we investigate the (computationally difficult) problem of finding which network nodes to select as proxies and we propose a simple method to address it. We demonstrate that the proposed method (i) guarantees that access delay stays below the given threshold, and (ii) significantly outperforms centralised and even distributed approaches, both in terms of access latency and in terms of maximum latency guarantees.