Bilateral Anomaly Detection Based on Edge Computing: Baseboard Management Controller

Anomaly detection on servers is a crucial technique with the increment of computing patterns in this period. To accommodate current demands, we propose a version of a bilateral detection scheme based on the baseboard management controller with machine learning to conduct compound anomaly detection and boost the upper bound. On the local part, the naive logic trees are deployed on the calculating module to make initial detection and carry out prompt actions maintaining the applicable working environment. Once the local anomaly alerts, the remote part is triggered to judge with the Gaussian Mixture Model, which ensures the anomaly type and proceeds with advanced steps. The bilateral detection system resorts to edge computing priority and scientifically allocates computing power to improve the level of anomaly detection on servers. The anomaly detection system’s effect herein is excellent in real-time demand and stability.