ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms
This addresses the problem of limited experimental infrastructure for ML research in Open RAN, benefiting researchers and developers in wireless networking, though it is incremental by providing a practical framework rather than a new ML method.
The paper tackles the slow progress in ML-based network automation for Open RAN by proposing ColO-RAN, a publicly-available large-scale testing framework with software-defined radios, enabling the design and evaluation of DRL-based closed-loop control, and demonstrates its effectiveness with three xApps on a network of 7 base stations and 42 users.
In spite of the new opportunities brought about by the Open RAN, advances in ML-based network automation have been slow, mainly because of the unavailability of large-scale datasets and experimental testing infrastructure. This slows down the development and widespread adoption of Deep Reinforcement Learning (DRL) agents on real networks, delaying progress in intelligent and autonomous RAN control. In this paper, we address these challenges by proposing practical solutions and software pipelines for the design, training, testing, and experimental evaluation of DRL-based closed-loop control in the Open RAN. We introduce ColO-RAN, the first publicly-available large-scale O-RAN testing framework with software-defined radios-in-the-loop. Building on the scale and computational capabilities of the Colosseum wireless network emulator, ColO-RAN enables ML research at scale using O-RAN components, programmable base stations, and a "wireless data factory". Specifically, we design and develop three exemplary xApps for DRL-based control of RAN slicing, scheduling and online model training, and evaluate their performance on a cellular network with 7 softwarized base stations and 42 users. Finally, we showcase the portability of ColO-RAN to different platforms by deploying it on Arena, an indoor programmable testbed. Extensive results from our first-of-its-kind large-scale evaluation highlight the benefits and challenges of DRL-based adaptive control. They also provide insights on the development of wireless DRL pipelines, from data analysis to the design of DRL agents, and on the tradeoffs associated to training on a live RAN. ColO-RAN and the collected large-scale dataset will be made publicly available to the research community.