Detection of DoS/DDoS attacks: the UBM and GMM approach

There are many different kinds of traditional techniques used for DoS/DDos attack detection, some of them include Artificial Intelligence, IDS, DPI, and most of them are well known and have remained unchanged during the last few years. In this work we implement two novel techniques called GMM and UBM, which are normally used in other scientific or engineering areas, to detect DoS/DDos cyberattacks using a real traffic dataset (CICIDS2017). Three experimental scenarios were implemented, including UBM, GMM, and a Random Forest alternative. The obtained results show new opportunities to use these novel methods in other approaches and explore new solutions to important problems like DoS/DDoS cyber-attack detection, which are closely related to a lot of services on the current Internet. The main goal of this work is to show the potential of UBM and GMM techniques in this simple problem and explore new applications in more complex scenarios.