.png){kind=icon}
.png){kind=icon}
Return
A FastAPI-Based Scalable ML Prediction Microservice for Real-Time Anomaly Detection in ONAP-Enabled Network Security
.png){kind=icon}
PRE
AI
S
W
DOI：10.1109/MAES.2025.3644327.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the era of increasingly complex and dynamic networks, traditional rule-based security systems struggle to keep up with emerging threats. Open Network Automation Platform (ONAP) offers a flexible orchestration and management framework that enables closed-loop automation. This article presents a scalable machine learning prediction microservice implemented with FastAPI, aimed at detecting anomalies in network traffic and integrate seamlessly with ONAP via the data collection, analytics, and events framework. Our solution supports real-time prediction through containerized deployment, achieving low-latency inference and facilitating automated response through ONAP's policy framework. The evaluation underscores the system's effectiveness in detecting command-and-control traffic patterns using packet-length features, demonstrating that lightweight packet-level features can enable accurate and efficient anomaly detection, and highlighting the framework's applicability to real-world deployments, particularly in next-generation communication environments, such as 6G and low-Earth orbit satellite systems.
Keywords:
ONAP
DCAE
Machine Learning
Anomaly Detection
Network Security
6 G
Journal
IF:
3.8
Papers:
2.1K
Citations:
2.5K
Organization
Cited Papers