A comprehensive detection and mitigation mechanism to protect SD-IoV systems against controller-targeted DDoS attacks
Behaylu Tadele Alemu,Alemu Jorgi Muhammed,Habtamu Molla Belachew,Mulatu Yirga Beyene
DOI: https://doi.org/10.1007/s10586-024-04660-8
2024-07-20
Cluster Computing
Abstract:Software-defined networking (SDN) has emerged as a transformative technology that separates the control plane from the data plane, providing advantages such as flexibility, centralized control, and programmability. This innovation proves particularly beneficial for Internet of Vehicles (IoV) networks, which amalgamate the Internet of Things (IoT) and Vehicular Ad Hoc Network (VANET) to implement Intelligent Transportation Systems (ITS). IoV provides a safe and secured vehicular environment by supporting V2V, V2I, V2S, and V2P. By employing an SDN controller, IoV networks can leverage centralized control and enhanced manageability, leading to the emergence of Software-Defined Internet of Vehicles (SD-IoV) as a promising solution for future communications. However, the SD-IoV networks introduces a potential vulnerability in the form of a single point of failure, particularly susceptible to Distributed Denial of Service (DDoS) attacks. This is because of the centralized nature of SDN and the dynamic nature of IoV. In this context, the SDN controller becomes a prime target for attackers who flood it with massive packet-in messages. To address this security concern, we propose an efficient and lightweight attack detection and mitigation scheme within the SDN controller. The scheme includes a detection module that utilizes entropy and flow rate to identify patterns indicative of attack traffic behavior. Additionally, a mitigation module is designed to minimize the effect of attack traffic on the normal operation, this is performed through analysis of payload lengths.The mitigation flow rule is set for specific traffic type if its payload is less than the threshold value to decrease the false positive rate. An adaptive threshold computation for all parameter values enhances the scheme's effectiveness. We conducted simulations using SUMO, Mininet-WiFi, and Scapy. We evaluated the system performance by using Mininet-wifi SDN simulation tool and Ryu controller for control plane. The system detects DDoS attack traffic within a single window by checking both entropy and flow rate simultaneously. The simulation results demonstrate the efficacy of our proposed scheme in terms of detection time, accuracy, mitigation efficiency, controller load, and link bandwidth consumption, showcasing its superiority compared to existing works in the field.
computer science, information systems, theory & methods