Distributed Denial of Service (DDoS) attacks are among the most critical cybersecurity threats, jeopardizing the stability of even the largest networks and services. The existing range of mitigation services predominantly filters at the edge of the Internet, thus creating unnecessary burden for network infrastructures. Consequently, we present IXP Scrubber, a Machine Learning (ML) based system for detecting and filtering DDoS traffic at the core of the Internet at Internet Exchange Points (IXPs) which see large volumes and varieties of DDoS. IXP Scrubber continuously learns DDoS traffic properties from neighboring Autonomous Systems (ASes). It utilizes BGP signals to drop traffic for certain routes (blackholing) to sample DDoS and can thus learn new attack vectors without the operator's intervention and on unprecedented amounts of training data. We present three major contributions: i) a method to semi-automatically generate arbitrarily large amounts of labeled DDoS training data from IXPs' sampled packet traces, ii) the novel, controllable, locally explainable and highly precise two-step IXP Scrubber ML model, and iii) an evaluation of the IXP Scrubber ML model, including its temporal and geographical drift, based on data from 5 IXPs covering a time span of up to two years.

In September 2019, 600 armed German cops seized the physical premise of a Bulletproof Hoster (BPH) referred to as CyberBunker 2.0. The hoster resided in a decommissioned NATO bunker and advertised to host everything but child porn and anything related to terrorism while keeping servers online no matter what. While the anatomy, economics and interconnection-level characteristics of BPHs are studied, their traffic characteristics are unknown. In this poster, we present the first analysis of domains, web pages, and traffic captured at a major tier-1 ISP and a large IXP at the time when the CyberBunker was in operation. Our study sheds light on traffic characteristics of a BPH in operation. We show that a traditional BGP-based BPH identification approach cannot detect the CyberBunker, but find characteristics from a domain and traffic perspective that can add to future identification approaches.