AI Application in Next Generation Programmable Networks

Abstract

Artificial intelligence (AI) algorithms can provide an effective solution for dynamic and automated network resource management in Software Defined Networking (SDN). In this contribution, we propose an auto-configuration enabler inside the next-generation Internet of Things (IoT) architecture proposed in the ASSIST-IoT project, for network resource allocation. The AI algorithm is responsible for controlling intent-based routing in an SDN network. This paper focuses on the problem of optimal intent switching between two designated paths using a Deep-Q-Learning approach based on an artificial neural network. The AI algorithm was trained to maximise the total throughput in the network and to use the network efficiently. The presented results confirm the validity of the applied AI approach to the problem of improving network performance in next-generation networks for economically and technically efficient implementation in the evaluation of IoT network systems.