China Communications · 2022 · 19 citations · 0 references
ComputingComputer SystemsCybersecurityInternet Of ThingsMachine LearningNetwork SecurityCongestion ControlDenial-of-service AttacksFlood MitigationIot SecurityReinforcement LearningSoftware-defined NetworkingAttack MitigationDeep Reinforcement LearningDenial Of Service Attack MitigationDistributed Denial Of ServiceFlooding AttacksNetwork Traffic EngineeringRényi EntropySoftware-defined Networking Security
Link flooding attack (LFA) is a fresh distributed denial of service attack (DDoS). Attackers can cut off the critical links, making the services in the target area unavailable. LFA manipulates legal low-speed flow to flood critical links, so traditional technologies are difficult to resist such attack. Meanwhile, LFA is also one of the most important threats to Internet of things (IoT) devices. The introduction of software defined network (SDN) effectively solves the security problem of the IoT. Aiming at the LFA in the software defined Internet of things (SDN-IoT), this paper proposes a new LFA mitigation scheme ReLFA. Renyi entropy is to locate the congested link in the data plane in our scheme, and determines the target links according to the alarm threshold. When LFA is detected on the target links, the control plane uses the method based on deep reinforcement learning (DRL) to carry out traffic engineering. Simulation results show that ReLFA can effectively alleviate the impact of LFA in SDN IoT. In addition, the rerouting time of ReLFA is superior to other latest schemes.