1. LightGBM-DDoS: Intelligent Model for Detecting Distributed Denial of Service Attacks in Software-Defined Networking.
- Author
-
Kabirat, Kareem M., Olaniyi, Aborisade D., and Adebukola, Onashoga S.
- Subjects
COMPUTER software ,DISEASE prevalence ,K-nearest neighbor classification ,CONTROLLERSHIP ,ALGORITHMS - Abstract
Software Defined Networking and other more sophisticated network infrastructure management have improved network programmability, flexibility, and cost-effectiveness over traditional approaches. By separating the network into layers, Software Defined Networking (SDN) technology simplifies network management. SDN networks are physically dispersed but logically centralized. Due to its fragmented structure, SDN is susceptible to different types of attacks. Distributed denial of service (DDoS) attacks is particularly problematic. The Controller is the main target of the attackers; once the Controller is affected, the entire network will shut down. Due to the prevalence of fake IP addresses, detecting DDoS attack traffic on SDN is more challenging. Therefore, effective and precise DDoS attack detection on the SDN is essential. This paper proposes a DDoS attack detection model on SDN leveraging the LightGBM-DDoS algorithm. The LightGBM-DDoS algorithm was benchmarked with XGBoost, Logistic Regression and K-Nearest Neighbours. The experimental results demonstrate that LightGBM-DDoS is superior to other approaches, with an accuracy rate, detection rate, F1-score, and error rate of 0.9972, 0.9940, 0.9970, and 0.0060, respectively. [ABSTRACT FROM AUTHOR]
- Published
- 2023