Assessment of Ensemble Bagging for Robust and Adaptive Identification of DDoS Attacks in Software-Defined Networks

Abstract

The increasing frequency and severity of Distributed Denial-of-Service (DDoS) attacks, which overload resources and block authorized access, pose a serious threat to the stability of modern networks. Robust and flexible security measures are needed to counter these threats. The incapacity of conventional protocols to adjust to changing threats is a problem that machine learning (ML) offers a workable solution to. Because ML algorithms can automatically detect fraudulent activity and improve accuracy, they are especially helpful in software-defined networks (SDN), where centralized controllers are very sensitive. This work explores machine learning-based methods for detecting DDoS attacks in SDN, with a focus on bagging decision tree algorithms and random forest ensemble technology. Performance was measured using accuracy, relative absolute error (RAE), and relative squared error (RRSE). The results show that bagging consistently outperforms random forest and yields higher detection efficiency. The proposed model unexpectedly outperforms existing benchmarks with an accuracy rate of 99.94%, suggesting that it could enhance SDN security.