Network Anomaly Flow Detection Framework Based on Collaborative Cloud-edge in the Smart Grid Network
Xiaolin Yang, Hao Zhang, Kai Yang, Yifei Lu · 2023
In recent years, the proliferation of Internet of Things devices in power grids has exposed these critical infrastructure systems to numerous malicious network attacks. These attacks are characterized by their extensive scale, scope, and potential for harm. Concurrently, smart power grids in China utilize poor-performing domestic devices, owing to policy mandates. Consequently, the centralized deep learning techniques commonly leveraged for anomaly detection on the Internet have limited applicability in power grids. To address this problem, we propose an online anomaly flow detection framework based on cloud-edge collaboration according to the actual distributed situation of power grids. High-performance cloud servers train online learning models, which are then deployed at substations for distributed detection. We evaluate the feasibility of our proposed solution through both offline and online testing.