Slicing Wi-Fi Networks for Differentiated IoT Service Provisioning

Network slicing is among the most wanted features of the 5G to address the diverse requirements of different classes of services, such as IoT applications. However, bringing the slicing to Wi-Fi networks is very challenging due to the lack of wireless virtualization supports in hardware. This paper proposes a new approach to achieve network slicing in Enterprise WiFi without virtualization techniques. Our approach is based on the dynamic association of user equipment to Wi-Fi access points to meet the requirements of differentiated IoT services. We model this technique as an optimization problem with the objective of maximizing the total throughput of the network while meeting the differentiated IoT service requirements. A heuristic algorithm based on the stable matching mechanism has been proposed to solve the optimization problem in near real-time. Furthermore, to enable a practical implementation, we advocate an online algorithm based on Reinforcement Learning which can mitigate the calculation in each time slot as done by the matching algorithm. Simulation results show that our solutions achieve the total throughput approximately to the optimum one and outperform the traditional RSSI method while guaranteeing the IoT slicing service requirements.