** 5G3E (5G End-to-End Emulation) dataset

The 5G3E (5G End-to-End Emulation) dataset contains monitoring metrics of a virtualized 5G infrastructure.

The dataset is created using real commercial cellular network traffic anonymized traces over a period of 14 days, injected over an emulated 5G platform including emulated radio access and 5G core network.

The possible usages of the dataset include, but are not limited to, learning and data analysis in relation to novel network automation architectures, protocols and algorithms.

** RELATED PUBLICATION

The dataset creation process and the dataset general characteristics are decribed in this paper:

Dung Chi Phung, Nour-El-Houda yellas, Salah Bin Ruba, Stefano Secci. An Open Dataset for Beyond-5G Data-driven Network Automation Experiments. 2022 1st International Conference on 6G Networking (6GNet), Jul 2022, Paris, 6-8 July 2022. hal-03698732. https://hal.archives-ouvertes.fr/hal-03698732

Any usage of the 5G3E dataset leading to any form of publication (private or public) must cite the above paper.

** DATASET DESCRIPTION

The dataset contains time-series system metrics, built at different sampling rates, related to the ob-servation of multiple resources involved in 5G network operation: radio, computing and network resources. The variety of collected features ranges from radio front-end metrics to physical server, from operating system to network function metrics. Few thousands of such features are collected, grouped by resource type and/or node type.

A detailed description of the employed 5G testbed at Cnam is given and of the v1 of the dataset are given in the paper above.

Additional versions with additional features (protocols, network functions, with edge computing elements, channel propagation models and events, etc) and setting (multipath transport, usage of WiFi, LiFi, Ethernet interfaces at the UE, etc) will be published further on.

** V1 DATASET DOWNLOAD

Version number: 1 Release date: June 26, 2022.

We include in this folder a sample of a 10' collection duration ; please do not download as bulk as github as download limitation policies preventing large downloads by many persons.

The 15-days dataset is fragmented in one zip file per day at these links:

Training set - day 1

Training set - day 2

Training set - day 3

Training set - day 4

Training set - day 5

Training set - day 6

Training set - day 7

Training set - day 8

Training set - day 9

Training set - day 10

Training set - day 11

Training set - day 12

Training set - day 13

Training set - day 14

Test set day

Write us if need any additional information. Email: Chi-Dung.Phung AT cnam DOT fr and Stefano.Secci AT cnam DOT fr

** ACKNOWLEDGEMENT

This dataset is created and maintained thanks to the support of the following public-funded research projects:

  • ANR CANCAN (Content and Context based Adaptation in Mobile Networks): http://cancan.roc.cnam.fr. Contract nb: ANR-18-CE25-0011.
  • H2020 AI@EDGE (A Secure and Reusable Artificial Intelligence Platform for Edge Computing in Beyond 5G Networks) project: https://aiatedge.eu. Grant nb. 101015922.
  • AMI-5G INFLUENCE project.