ET-RF based Model for Detection of Distributed Denial of Service Attacks

Abstract

Distributed Denial of Service (DDoS) attack is a type of network attack that can be launched from multiple sources to bring the network down. Several detection algorithms have been adopted to diagnose Distributed Denial of Service attacks. In this paper, the authors proposed an ET-RF (Extra Tree-Random Forest) model on CICDDoS2019 dataset to detect DDoS attacks. The system has been tested in two scenarios on CICDDoS2019 dataset. In scenario 1 the performance of different classifiers Random Forest, Decision Tree and KNN (K-Nearest Neighbor) have been evaluated. Analysis using ROC Curve gives 99% accuracy for Random Forest with Extra Tree feature selection on complete dataset. In scenario 2 the authors explored tests with different types of DDoS attacks. Since, all the attacks are analyzed independently and recall, f-1 score and precision close to 99% are achieved using this model.