LFA-Cleaner: Mitigating Link Flooding Attacks in Software-Defined Networks
Jirun Xie, Liang Wei, Shan Jing, Chuan Zhao, Zhenxiang Chen · 2023
Link Flooding Attack (LFA) is a network attack that targets backbone or bottleneck links in a network, which is difficult to detect and can cause significant damage. In this paper, we propose a new framework for detecting LFA, LFA-Cleaner, which exploits the feature of Software Defined Network (SDN) to have a global view to find network bottlenecks and deploy lightweight load balancing algorithms to relieve the pressure. The Random Forest algorithm and Gradient Boosting Decision Tree algorithm in ensemble learning are also used in combination with the mitigation module proposed in this paper to achieve the detection and defense against LFA. Finally, we experimentally verify the effectiveness of our approach.