Slicing Optimization based on Machine Learning Tool for Industrial IoT 4.0

Abstract

Industry 4.0 is considered as a very promising paradigm for Industrial Internet of Things (IIoT) that will significantly impact current industries and the construction of upcoming ones due to its various use cases. The latter have heterogeneous quality of service (QoS) requirements which imposes important challenges in enabling these applications over a single IIoT infrastructure. In this paper, we propose an SDN-based architecture for Industry 4.0 as well as a dynamic slicing admission and resource reservation method based on online machine learning tools to provide flexibility in managing network resources while avoiding performance degradation of urgent IIoT traffic with network slicing. Simulation results, implemented over NS3 network simulator, highlights the efficiency of our proposed method in avoiding resources starvation and providing QoS for devices by respecting the defined delay thresholds and decreasing energy consumption.