An Anomaly Detection Method for Data Center Network Traffic Based on Dynamic Graph Convolutional Networks

Key Points

• This research aims to develop a new method for detecting anomalies in data center network traffic using dynamic graph convolutional networks.
• Utilized dynamic graph convolutional networks for modeling network traffic data.
• Implemented anomaly detection algorithms to identify irregular traffic patterns.
• Conducted performance evaluations against existing approaches.
• Achieved a significant reduction in false positive rates (FPR) compared to traditional methods.
• Improved detection accuracy with dynamic graph techniques, enhancing overall network security.