Enhanced Anomaly Detection Framework for 6G Software-Defined Networks: Integration of Machine Learning, Deep Neural Networks, and Dynamic Telemetry

Abstract

The advent of 6G networks ushers in a new era of intelligent network management, necessitating robust security measures to safeguard against emerging threats. This paper presents a comprehensive framework for anomaly detection tailored specifically for 6G Software-Defined Networks (SDNs), leveraging innovative ML), (DL), and dynamic telemetry techniques. The proposed framework, termed Anomaly Detection System for 6G SDNs, integrates ensemble learning (EL) algorithms and deep neural networks (DNNs) to detect anomalies within network traffic. Beginning with the preprocessing and feature selection stages, the proposed system employs an amalgam EL method to enhance the efficacy of anomaly detection. Datasets including CICDDOS2019, NSL KDD, CIC_IDS2017, and NB2015 undergo dimensionality reduction and feature subset determination to optimize performance. Furthermore, dynamic telemetry is seamlessly integrated into the proposed, enabling real- time monitoring and adaptive response mechanisms within SDN environments. By harnessing the flexibility and programmability of SDNs, the framework ensures a proactive defense against evolving threats, bolstering the security posture of 6G networks. Experimental evaluations demonstrate the effectiveness of ADS6SDN across diverse datasets, achieving high accuracies while minimizing false alarm rates. In conclusion, integrating ML, DL, and dynamic telemetry within the proposed approach offers a potent solution for enhancing the security and responsiveness of 6G SDNs. By leveraging the inherent advantages of SDN architectures, the framework not only fortifies network defenses against emerging threats but also ensures adaptability to the budding scenario of next-generation telecommunications.