Scalable DDoS Attack Mitigation in SDN via Real-Time Adaptive Machine Learning Model
Raghul Napoleon, Ramu Naresh · 2025
Current systems employed for DDoS attack detection in SDN environments with high false positive rates and slow responses are inefficient in dealing with changed attack patterns. This is due to the static nature of the traditional detection mechanisms that cannot adapt in real time to the threats. Toward this end, the proposed solution incorporates an ensemble online machine learning model consisting of three XGBoost, Support Vector Machine, and Decision Tree algorithms. The model of the said arrangement continuously learns from the network traffic data for improved detection accuracy and changes in the strategy of attacks over time. The overall aim here is to make DDoS detection and mitigation as powerful as possible with the help of SDN’s ability to provide real-time programmability. The system is most beneficial in terms of faster decision making, less false positives, and efficient rerouting of traffic during an attack. The key advantages include the real-time adaptability, scalability, and the improved resilience of the network against DDoS attacks; hence, highly suited for the modern, dynamic network environments.