My research focuses on scalable and resource-efficient detection of volumetric Distributed Denial-of-Service (DDoS) attacks using machine learning. I explore novel traffic monitoring approaches—Traffic Aggregates and Traffic Images—that enable fast and fine-grained detection without relying on microflow-level analysis. These approaches ensure fixed memory and computational requirements while preserving key detection-relevant features such as IP and port distributions. To address the challenge of labeled training data, I develop autonomous, feedback-driven labeling methods that eliminate manual annotation and adapt to evolving traffic patterns. Overall, my work aims to make ML-based DDoS detection deployable at scale in real-world high-speed networks.

• KIWI - Artificial intelligence in secure web infrastructures with digital identity management
• bwNET2.0 - A more powerful and flexible network for Baden-Württemberg