Adaptive end-to-end monitoring framework for heterogeneous 5G and beyond networks

Abstract

5G technology has ushered in complex multi-domain environments that demand dynamic and robust monitoring solutions. Traditional network monitoring strategies often fail to address the unique challenges of 5G networks, such as managing high data volumes, diverse service requirements, and cross-domain interactions across heterogeneous technologies. In this paper, we present an innovative end-to-end monitoring framework for 5G and beyond networks designed to enhance failure detection and localization capabilities, as well as overall performance evaluation across diverse 5G deployments. Our framework leverages statistical learning techniques to efficiently and adaptively analyze network data, focusing on specific traffic sub-populations that reflect immediate monitoring needs. The proposed monitoring system is designed to be programmable and application-sensitive, allowing on-the-fly configuration changes that are essential for multi-domain operations. By integrating flow-based measurements with intelligent sampling methods, our system significantly reduces the resource footprint traditionally required for comprehensive data collection and analysis. We have implemented and validated our framework using the ns-3 5G-LENA simulator. This approach enables us to evaluate the system's performance in realistic 5G scenarios and demonstrate its effectiveness across various network conditions and configurations, addressing the challenge of limited access to commercial 5G deployments. Preliminary results from our simulations demonstrate the framework's potential to remarkably improve network reliability, performance insights, and operational efficiency across heterogeneous 5G environments. Our approach facilitates more precise and scalable network management, setting the stage for adaptive monitoring solutions as 5G and beyond networks' demands evolve.