An Information Entropy and Ensemble Learning Approach for DR-DOS Detection within Aviation Networks

Fifth Generation (5G) architecture brings a new wave of innovation into the development of the future aviation environment. However, the deployment of Internet of Things (IoT) infrastructure weakens the current closed loop systems deployed within the airport operational domain; caused by the expansion of attack surfaces due to an increased number of endpoints. We propose a solution to detect multiple Distributed Reflective Denial-of-Service (DRDoS) attack types utilizing stacked ensemble structure of Bayesian Statistics, J48 (C4.5) Decision Trees and Multi-Layer Perception algorithms, to identify DRDoS attacks. The proposed solution gives a 94.1% accuracy on the CIC-2019 dataset.