Design and Implementation of an Anomaly Detection and Traffic Prediction Mechanism

With the growing deployment of the Smart Integration Routing System (SIRS), ensuring system security has become a pressing challenge. Internal failures and external traffic attacks significantly impact system robustness, while the complexity of SIRS routing devices complicates anomaly detection and root cause analysis. To address these challenges, this paper proposes a comprehensive anomaly detection and prediction framework for SIRS traffic data, focusing on the critical network traffic. Furthermore, the Knowledge Graph Convolutional Network (KGCN) is employed for root cause localization, enabling the identification of complex, non-Euclidean relationships between anomalies and their causes. This study provides a robust and scalable solution for anomaly detection, prediction, and root cause localization in the Smart Integration Routing System, significantly improving its security, maintainability, and resilience against dynamic threats.