SmartNIC-Accelerated Intrusion Detection with Nested Graph Neural Networks
Details
Abstract: This talk presents a hardware–software co-designed intrusion detection system that leverages SmartNIC acceleration and Nested Graph Neural Networks to enable high-throughput, low-latency security analytics in modern networks.
Speaker: Rana Abu Bakar received his PhD in Emerging Digital Technologies from Scuola Superiore Sant'Anna and is working as a researcher in CNIT, Italy. Currently working on SmartNIC/DPU-accelerated networking and AI-driven security systems. His research focuses on high-speed packet processing, hardware offloading, and graph-based learning for intrusion and anomaly detection. He has published and presented work at leading IEEE venues and is actively involved in EU research projects on trustworthy and high-performance network security.