dApps: Enabling Real-Time AI-Based Open RAN Control

In this repository we configure an OpenAirInterface 5G gNB and a dApp for real-time RAN control.

There are two main modules composing this frameworks:

• A custom library in Python for dApps
• A custom version of OAI (soon to be updated and hopefully merged with the original branch)

For a more detailed description of what are the dApps, please refer to our recent architecture paper and this page.

Usage

A complete tutorial on dApps can be found on the OpenRAN Gym website.

In this page we discuss the configuration parameters that can be use to perform experiments over the air (OTA).

Configuration for OAI

For every configuration of the E3 interface, make sure that do_SRS flag is set to 0.

System configuration for OTA

This part of the guide is inteded to be only if running with --ota and with Unix Domain Sockets (UDSs). If we need the dApp to run in a different user than OAI (e.g., $USER vs root) we need to create a specific unix users group called dapp and we assign root and user to this group to enable shared UDS through a dedicated folder:

sudo groupadd dapp
sudo usermod -aG dapp root
sudo usermod -aG dapp $USER

# check the groups
groups root $USER

Folder creation (path is not important, but it should be the same in OAI and the dApp)

mkdir -p /tmp/dapps
sudo chown :dapp /tmp/dapps
sudo chmod g+ws /tmp/dapps

If the dApp is not run from the root user, a new folder should be included called logs

# In the root of your project
mkdir logs

If you use the dApp concept and/or the framework to develop your own dApps, please cite the following paper:

@ARTICLE{lacava2025dApps,
author = {Andrea Lacava and Leonardo Bonati and Niloofar Mohamadi and Rajeev Gangula and Florian Kaltenberger and Pedram Johari and Salvatore D'Oro and Francesca Cuomo and Michele Polese and Tommaso Melodia},
title = {dApps: Enabling Real-Time AI-Based Open RAN Control},
url = {https://arxiv.org/pdf/2501.16502},
pages = {1-31},
year = {2025}
}