A flow-level taxonomy and prevalence of brute force attacks

Abstract

Online brute force and dictionary attacks against network services and web applications are ubiquitous. We present their taxonomy from the perspective of network flows. This contributes to clear evaluation of detection methods and provides better understanding of the brute force attacks within the research community. Next, we utilize the formal definitions of attacks in a long-term analysis of SSH traffic from 10 gigabit university network. The results shows that flow-based intrusion detection may profit from traffic observation of the whole network, particularly it can allow more accurate detection of the majority of brute-force attacks in high-speed networks. © 2011 Springer-Verlag.