This white paper explores how AI and machine learning can strengthen network security against distributed denial-of-service (DDoS) attacks. Traditional methods like ACLs, rate limiting, and blackholing are reactive and resource-intensive. By contrast, the paper demonstrates how a random forest classifier trained on datasets such as CIC-IDS2017 and KDD Cup 1999 can predict and classify malicious traffic with high accuracy (>99%). Implemented on AMD Versal adaptive SoCs, the model achieves millions of inferences per second, offering scalable, low-latency protection. Future extensions include inline FPGA-based flow processing and deep learning models for broader anomaly detection.

Join for free to read