Intelligence and Learning in O-RAN for Data-Driven NextG Cellular Networks

Bonati, Leonardo; D’Oro, Salvatore; Polese, Michele; Basagni, Stefano; Melodia, Tommaso

doi:10.1109/mcom.101.2001120

Cited by 143 publications

(65 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Differently from these papers, we analyze the performance of DRL agents with a closed loop, implementing the control actions on a software-defined testbed with an O-RAN compliant infrastructure to provide insights on how DRL agents impact a realistic cellular network environment. Finally, [6,7] consider ML/DRL applications in O-RAN, but provide a limited evaluation of the RAN performance without specific insights and results on using ML.…”

Section: Related Workmentioning

confidence: 99%

“…Last, data-driven modeling will exploit recent developments in ML and big data to enable realtime, closed-loop, and dynamic decision-making based, for instance, on Deep Reinforcement Learning (DRL) [5]. These are the very same principles at the core of the Open RAN paradigm, which has recently gained traction as a practical enabler of algorithmic and hardware innovation in future cellular networks [6][7][8].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms

Polese¹,

Bonati²,

Basagni³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Cellular networks are undergoing a radical transformation toward disaggregated, fully virtualized, and programmable architectures with increasingly heterogeneous devices and applications. In this context, the open architecture standardized by the O-RAN Alliance enables algorithmic and hardware-independent Radio Access Network (RAN) adaptation through closed-loop control. O-RAN introduces Machine Learning (ML)-based network control and automation algorithms as socalled xApps running on RAN Intelligent Controllers. However, 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 largescale dataset will be made publicly available to the research community.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms

Polese¹,

Bonati²,

Basagni³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Accordingly, the use of COTS equipment from any vendor will potentially facilitate self-coordination capabilities because the network control parameters (NCPs) will be similar among various network elements. This can also reduce the parameter types to be controlled, and alleviate multi-vendor compatibility issues [12]. On the downside, different SONFs can access the same database and consequently compromise each other's security.…”

Section: B Future Research Directionsmentioning

confidence: 99%

Hybrid Self-Organizing Networks: Evolution, Standardization Trends, and a 6G Architecture Vision

Chaoub¹,

Mämmelä²,

Martínez-Juliá³

et al. 2021

Preprint

View full text Add to dashboard Cite

Self-organizing networks (SONs) need to be endowed with self-coordination capabilities to manage the complex relations between their internal components and to avoid their destructive interactions. Existing communication technologies commonly implement responsive self-coordination mechanisms that can be very slow in run-time situations. The sixth generation (6G) networks, being in their early stages of research and standardization activities, open new opportunities to opt for a design-driven approach when developing self-coordination capabilities. This can be achieved through the use of hybrid weakly coupled SON designs. In this article, we review the history of SONs including the inherent self-coordination feature. We then delve into the concept of hybrid SONs (H-SONs), and we summarize the challenges, opportunities, and future trends for H-SON development. We provide a comprehensive collection of standardization activities and recommendations, discussing the key contributions and potential work to continue the evolution and push for a wide adoption of the H-SON paradigm. More importantly, we propose that H-SONs must be weakly coupled networks, i.e., the various feedback loops must be almost isolated from each other to improve the stability and to avoid chaotic situations. We finally conclude the paper with the key hints about the future landscape and the key drivers of 6G H-SONs.

show abstract

“…Telecom operators will deploy open source custom solutions on a "white-box" agnostic Radio Access Network (RAN) where services, e.g., the base stations, will be instantiated on-demand [1,22]. This not only endows the network with flexibility by design, but it also allows real-time optimization of the users' service based on the current network conditions and traffic demand [16,23]. Although open-source approaches bring unprecedented performance improvements, they require a reliable development environment and at scale evaluation before they are deployed on the commercial infrastructure.…”

Section: A Cellular Networkingmentioning

confidence: 99%

“…1 Origi- With its 256 SDRs and 128 remotely-accessible compute nodes and GPUs, Colosseum provides the capabilities to test full-protocol stack solutions at scale with real hardware devices and in emulated-yet realistic-environments with complex channel interactions (e.g., path loss, fading, multipath). Besides its experimentation capabilities, Colosseum can be used as an AI playground and wireless data factory to create large-scale datasets and train/test solutions in a safe and controlled environment [15,16]. We provide an overview of the architectural components and emulation capabilities of Colosseum in Sections II and III.…”

Section: Introductionmentioning

confidence: 99%

Colosseum: Large-Scale Wireless Experimentation Through Hardware-in-the-Loop Network Emulation

Bonati¹,

Johari²,

Polese³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Colosseum is an open-access and publicly-available large-scale wireless testbed for experimental research via virtualized and softwarized waveforms and protocol stacks on a fully programmable, "white-box" platform. Through 256 state-of-theart Software-defined Radios and a Massive Channel Emulator core, Colosseum can model virtually any scenario, enabling the design, development and testing of solutions at scale in a variety of deployments and channel conditions. These Colosseum radiofrequency scenarios are reproduced through high-fidelity FPGAbased emulation with finite-impulse response filters. Filters model the taps of desired wireless channels and apply them to the signals generated by the radio nodes, faithfully mimicking the conditions of real-world wireless environments. In this paper we describe the architecture of Colosseum and its experimentation and emulation capabilities. We then demonstrate the effectiveness of Colosseum for experimental research at scale through exemplary use cases including prevailing wireless technologies (e.g., cellular and Wi-Fi) in spectrum sharing and unmanned aerial vehicle scenarios. A roadmap for Colosseum future updates concludes the paper.

show abstract

Intelligence and Learning in O-RAN for Data-Driven NextG Cellular Networks

Cited by 143 publications

References 9 publications

ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms

ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms

Hybrid Self-Organizing Networks: Evolution, Standardization Trends, and a 6G Architecture Vision

Colosseum: Large-Scale Wireless Experimentation Through Hardware-in-the-Loop Network Emulation

Contact Info

Product

Resources

About