ML-Based Delay Attack Detection and Isolation for Fault-Tolerant Software-Defined Industrial Networks.

Area of Science:

• Cyber-Physical Systems Security
• Network Resilience
• Machine Learning Applications

Background:

• Traditional security struggles with intelligent cyber-physical system assaults.
• Software-defined networks (SDNs) in industrial settings face high resource consumption and routing inefficiencies.
• Existing resilience schemes like SDN-RM are vulnerable to specific delay attacks.

Purpose of the Study:

• To develop a real-time delay attack detection and isolation scheme for fault-tolerant software-defined industrial networks.
• To enhance the resilience of the SDN-RM scheme against link layer discovery protocol (LLDP) packet delay attacks.
• To improve network performance metrics such as success rate and throughput.

Main Methods:

• Developed a machine learning (ML)-based attack detection and isolation mechanism.
• Integrated an ML mechanism with a route-handoff mechanism to create the DA-DIS scheme.
• Focused on detecting attacks that delay LLDP packets by compromising OpenFlow switches.

Main Results:

• The proposed DA-DIS scheme effectively detects and isolates malicious switches.
• DA-DIS prevents compromised switches from being included in network routes.
• The scheme significantly increases network resilience, success rate, and throughput.

Conclusions:

• Integrating ML with network resilience solutions is effective for identifying malicious network components.
• The DA-DIS scheme offers a robust solution for enhancing the security and performance of industrial SDNs.
• The proposed method addresses the limitations of traditional security mechanisms in dynamic industrial network environments.