A Review on Deep Reinforcement Learning for the management of SDN and NFV in Edge-IoT
Ricardo S. Alonso, Javier Prieto, Fernando De la Prieta, Sara Rodrı́guez, Juan M. Corchado
Abstract
The Internet of Things (IoT) represents billions of devices collecting valuable information through different sensors to be transferred to the Cloud, where it is stored and processed to infer valuable knowledge. In this regard, Edge Computing (EC) architectures allow providing shorter service response times and reducing the cost of processing IoT data in the Cloud. Moreover, Network Function Virtualization (NFV) provide mechanisms to reduce costs when sharing physical network resources transparently by different user entities. Software-Defined Networks (SDN), closely related to the NFV, are also used to manage virtual networks in a flexible and dynamic way. However, different applications on shared IoT networks may demand different Quality of Service (QoS). Therefore, intelligent mechanisms and algorithms are needed to optimize virtual data flows in networks, such as Deep Reinforcement Learning techniques. This papers presents a review of the existing approaches on DRL for the management of SDN/NFV in Edge-IoT scenarios.