Mitigation of Distributed Denial of Service (DDoS) Attacks in Software-Defined Networking Using Machine Learning Classifiers
The centralised controller of software-defined networking (SDN) is an attractive target for distributed denial of service (DDoS) attacks, yet many machine-learning studies evaluate detection on randomly split data and do not implement mitigation. This study designed a detection and mitigation module for the Ryu controller that classified traffic every three seconds using nine flow-statistics features and, upon detection, installed OpenFlow drop or meter rules at the ingress port. Decision Tree, Random Forest, k-Nearest Neighbours, Support Vector Machine, Naive Bayes and XGBoost were compared on the InSDN dataset and on Mininet testbed traffic containing TCP SYN, UDP and ICMP floods, using a time-ordered split. XGBoost and Random Forest performed best, with F1-scores of 99.1% and 98.9% on the testbed, and XGBoost required only 0.8 ms per classification. With mitigation, attacks were detected within 4.6 s on average, benign throughput recovered to 95% within about 11 s, and peak controller CPU utilisation fell from 96% to 38%. Low-rate attacks remained the most difficult case, with a detection rate of 92.8%. Lightweight tree ensembles integrated in the controller therefore provided effective real-time DDoS protection.
Authors
- Hendra Wijaya
- Nurul Hidayah
Publication Details
- Journal
- Journal of Information Technology computer science and Electrical Engineering
- Published
- 2026-11-18
- DOI
- https://doi.org/10.61306/w1zpav33
- Primary Topic
- Network Security and Intrusion Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00