We release a new version of the 5G3E dataset, moving from its foundational first edition to a significantly more powerful, cloud-native Version 2.
A curated dataset with end-to-end cloud native 5G stack emulation
5G End-to-End Emulation (5G3E): simplified setting
From Foundation to Innovation: The 5G3E Journey
The first edition of the 5G3E dataset was designed to provide a comprehensive look at the multi-layered nature of 5G operations. It captured thousands of time-series system metrics across three vital pillars: Radio, Computing, and Network resources. By monitoring everything from radio front-end metrics to physical server performance and operating system health, Version 1 gave researchers a holistic view of a 5G network.
The second edition of the 5G3E dataset represents a massive leap forward. While it maintains the rich time-series variety of its predecessor, it is built upon a completely updated, fully containerized testbed. Here are the key improvements that make V2 a game-changer for researchers and engineers:
- A Shift to 5G Standalone (SA) Architecture 🏗️
- End-to-End Containerization 🐳
- Updated & Flexible Software Stack ⚙️
- Expanded Metrics & Standardized Compliance 📊
Whether you are training machine learning models to detect network anomalies or other tasks related to state assessment of 5G network infrastructure, the 5G3E V2 dataset provides the granular, end-to-end data you need. For a deeper dive, be sure to check out the repository https://github.com/cedric-cnam/5G3E-dataset/tree/main
This work is made possible through the dedicated support of several public-funded research projects:
- ANR TREES: Focusing on energy-efficient distributed learning for 6G. Learn more
- ANR COCO 5G: Advancing the future of 5G connectivity. Learn more
- PIIEC ME/CT PART. Learn more
