Real-time processing of cybersecurity system data for attacker profiling

Usage of cybersecurity tools entails an enormous amount of data that brings the possibility of different approaches to the processing of cybersecurity data. This paper discusses the profiling of attackers, which, in practice, can help in managing cybersecurity events. The main goal of the research is to perform attackers' profiling as close as possible to real-time processing. The paper outlines the basic idea of real-time attacker profiling. We use stream processing. Within the system, we profile attackers into seven profiles or mark them as outliers if they do not fall into any of the known profiles. The paper also deals with the dynamic profiling model update and the difference of the calculated model using the original non-real-time model.