AI-Driven Zero-Touch Containerized Network Functions Optimization for 6G Networks Using Network Data Analytics Components

Abstract

Emerging technologies and use cases, such as telepresence, cobots or massive twinning, bring a new level of demand on performance to future 6G networks. To meet the increasingly strict network requirements of the previously mentioned use cases, networks must go beyond traditional 5G capabilities, enhancing flexibility and resource optimization to support these forthcoming applications, while also reducing the carbon footprint of the virtualized network appliances.In this paper we investigate the integration of AI-driven Zero-Touch technologies in containerized Virtual Network Functions (VNFs) for future 6G networks. For achieving the presented results we leverage on a Kubernetes-Based emulated cluster, a Machine Learning (ML)-enabled Network Data Analytics Function (NWDAF), and both technologies combined to enhance performance and energy efficiency in a Cloud-Native and open-source 5G network. With this approach we aim at optimizing network load management in real-time and predicting network behavior to automate resource allocation effectively, thereby reducing human intervention and ensuring robust, scalable network architectures.