CLEVER Project delivered cutting-edge research in high-performance network scanning, hardware acceleration, and AI-driven threat detection:
Network Scanning & Performance
- SmartNIC + DPDK Integration: Reduced CPU utilization by ~40% using offload engines for efficient packet processing.
- Advanced Scanning Techniques: Protocol-specific probes and evasive methods enhance detection while bypassing protection systems.
- Core Network Deployment: Minimized probe path hops (2-4 vs. 4-8), reducing bandwidth waste and packet loss.
- 100 GbE Traffic Handling: Leveraged NVIDIA A100 GPUs and DPDK for high-speed packet processing on commodity hardware.
- Vulnerability Detection: Identified critical exploits (Heartbleed, Shellshock, EternalBlue) with high accuracy.
Hardware & Theoretical Optimization
- User/Kernel Space Management: Optimized resource utilization, improving scanning speed.
- Mathematical Modeling: Applied queuing theory and Markov chains to model latency in DPDK-based scanners.
- Hardware Acceleration: Achieved orders-of-magnitude improvements in GPU processing and packet generation.
AI-Powered Security for 6G
- Unsupervised Deep Learning: Combined autoencoders and Gaussian Mixture Models (GMM) to detect DoS attacks (Hulk, Goldeneye) with F1-scores up to 97.3%.
- Collective Network Intelligence: Paving the way for autonomous, real-time threat mitigation in 6G networks.
🔗 Explore the technical details: