“Autoencoder based Anomaly detection” ― a network automation platform to discover and characterize anomalies in real-time.
“Autoencoder based Anomaly detection” is a network automation framework which aims to learn nominal operating conditions of a softwarised network service and characterise anomalies in real-time, while offering a compact system state representation called radiography. This representation can provide operational teams with a real-time insight on anomalies at physical and virtualised layers. This tool can, for example, detect real-time anomalies of different nature on a containerised vIMS (virtual IP Multimedia Subsystem) service managed by Kubernetes.
Available at github.com/Orange-OpenSource/Autoencoder-Based-Anomaly-Detection.