Search papers, labs, and topics across Lattice.
This paper introduces MLDAS, a framework that dynamically selects the most appropriate machine learning algorithm for intrusion detection in Software-Defined Networks (SDNs) based on real-time network traffic characteristics. The system analyzes traffic-type-based metrics to define classification rules and adaptively chooses ML models, addressing the limitations of static SDN-based attack detection. Results demonstrate improved detection robustness and operational feasibility under varying network conditions by mitigating overfitting and underfitting through hyperparameter tuning.
Forget static defenses: MLDAS dynamically swaps intrusion detection algorithms in SDNs to maintain peak performance under shifting network conditions.
Network security is a critical concern in the digital landscape of today, with users demanding secure browsing experiences and protection of their personal data. This study explores the dynamic integration of Machine Learning (ML) algorithms with Software鈥怐efined Networking (SDN) controllers to enhance network security through adaptive decision mechanisms. The proposed approach enables the system to dynamically choose the most suitable ML algorithm based on the characteristics of the observed network traffic. This work examines the role of Intrusion Detection Systems (IDS) as a fundamental component of secure communication networks and discusses the limitations of SDN鈥恇ased attack detection mechanisms. The proposed framework uses adaptive model selection to maintain reliable intrusion detection under varying network conditions. The study highlights the importance of analyzing traffic鈥恡ype鈥恇ased metrics to define effective classification rules and enhance the performance of ML models. Additionally, it addresses the risks of overfitting and underfitting, underscoring the critical role of hyperparameter tuning in optimizing model accuracy and generalization. The central contribution of this work is an automated mechanism that adaptively selects the most suitable ML algorithm according to real鈥恡ime network conditions, prioritizing detection robustness and operational feasibility within SDN environments.
