6G-SANDBOX

Outcomes:

• A disaggregated RAN architecture has been delivered and tested across two servers with RIC/CU/DU and Open5GS components to ensure modularity and scalability. The delivered RAN architecture leverages full 10G connectivity between CU and DU, with front-haul links carefully monitored for packet loss. The configuration supports flexible deployment, making it a robust foundation for performance benchmarking and future beyond 5G innovations.
• Two xApps for power control of the RUs in a multi-cell environment have been developed. The ML models have been tested in an NDT simulated environment and integrated in the RI, along with the prototype reference CM. The CM functionality has been validated, showcasing its capability to detect direct conflicts across different power control actions and resolve them based on UE-centric or EE-centric policy.

Impact of the results:

• For the 6G-SANDBOX consortium: the delivered architecture enables the collection of measurements and data from the radio (RU/DU) that can be exploited by diverse 6G scenarios, thereby facilitating experimenters to conduct testing on 6G services towards future innovations. Moreover, the implementation of the ML-assisted algorithms in the O-RAN environment is included, showcasing the adoption of automated radio control.
• For the REACT-6G consortium partners: the involved partners are planning to be commercially active in 6G technologies, developing and integrating additional RAN equipment and components, as well as ML algorithms residing in xApps in the framework of O-RAN architecture. In this context, FDI, ACC and SRS have already established relationships with members of 6G-SANDBOX’s consortium, aiming to jointly develop O-RAN solutions for 6G networks with high commercial interest for the international markets.
• Environmental Impact is implicitly achieved through the objectives of the delivered ML algorithms for power control. The ML-assisted energy efficiency operation of the radio units was