Distributed denial of service detection and mitigation in software-defined networking-enabled software-defined wide...

Area of Science:

• Computer Science
• Network Security
• Machine Learning

Background:

• Software-Defined Wide Area Networks (SD-WAN) offer flexibility but introduce security vulnerabilities, especially against Distributed Denial of Service (DDoS) attacks.
• SD-WAN controllers are critical targets for DDoS attacks, necessitating robust defense mechanisms.

Purpose of the Study:

• To propose and evaluate an advanced security framework for SDN-enabled SD-WAN to detect and mitigate DDoS attacks.
• To develop an adaptive machine learning model capable of accurately identifying both high-rate and low-rate DDoS attacks.

Main Methods:

• Leveraged machine learning algorithms (Random Forest and Decision Tree) integrated within an adaptive framework for DDoS attack detection.
• Utilized a QT-PCA preprocessing pipeline for dimensionality reduction and PACKET_IN event-triggered mitigation for dynamic response.
• Trained and tested the model using a dataset collected from experimental environments.

Main Results:

• Achieved high accuracy in detecting DDoS attacks: 99.97% for high-rate and 99.96% for low-rate attacks.
• Demonstrated the framework's ability to adapt to dynamic network conditions and enhance network resilience.
• The QT-PCA pipeline effectively reduced dimensionality without compromising performance.

Conclusions:

• The proposed security framework provides robust and accurate DDoS attack detection for SDN-enabled SD-WAN.
• The adaptive machine learning model enhances network security and resilience against sophisticated threats.
• The approach optimizes network performance, proving practical for real-world deployment.