Towards Sustainability in 6G and beyond: Challenges and Opportunities of Open RAN

Hamed Ahmadi, Mostafa Rahmani Ghourtani, Swarna Bindu Chetty, Eirini Eleni Tsiropoulou, Huseyin Arslan, Mérouane Debbah, Tony Q. S. Quek

Abstract

While 6G aims to be 100x more energy-efficient than 5G, the explosion in data demand threatens to drive total energy consumption and e-waste to unsustainable levels. This paper argues that Open RAN (O-RAN) is the essential architectural shift needed to achieve true sustainability, using softwarisation, AI-driven resource management, and "white-box" hardware to break the cycle of proprietary waste.

Introduction: The Efficiency Paradox

The telecommunications industry faces a looming "Efficiency Paradox." Even as 6G improves energy efficiency (bits per Joule), the sheer volume of anticipated traffic and the massive hardware upgrades required for new services—such as Integrated Sensing and Communication or Digital Twins—could lead to a net increase in carbon emissions and electronic waste.

Traditional Radio Access Networks (RAN) are "black boxes"—proprietary, tightly coupled hardware and software from a single vendor. When a new generation arrives, the old hardware is often scrapped because it cannot be repurposed. This paper addresses the need for a holistic definition of sustainability encompassing economic (vendor diversification), environmental (e-waste and renewables), and social (global accessibility) dimensions.

The O-RAN Sustainability Framework

The authors propose Open RAN as the "recipe" for 6G sustainability, focusing on three technological pillars:

  1. Softwarisation & Cloudification: By decoupling network functions from hardware, O-RAN allows services to run on General-Purpose Processors (GPPs). This extends hardware lifespans and allows for "recycling" compute power across different parts of the network.
  2. RAN Intelligent Controllers (RIC): The architecture utilizes xApps and rApps—specialized software plugins that use AI/ML to dynamically optimize the network. For example, an energy-saving xApp can hand over users to neighboring cells and put underutilized Radio Units (RUs) into deep sleep.
  3. Cell-Free (CF) Massive MIMO: The authors simulate a "Cell-Free" architecture where users are served by a dispersed array of access points rather than a single tower, managed via O-RAN principles and efficient fronthaul protocols like eCPRI.

Sustainable Architecture and AI

Sustainability in O-RAN extends beyond the electricity bill. By using "white-box" hardware, operators can upgrade software without discarding physical servers, directly tackling the global e-waste crisis. Furthermore, O-RAN’s disaggregated nature allows the "brains" of the network (the RIC) to be placed at the edge, where it can be directly powered by local renewable sources.

Vision for a sustainable O-RAN ecosystem integrating green energy and AI-driven optimization.

However, AI itself presents a challenge. Training large models is energy-intensive. The authors advocate for "Energy Sufficiency"—encouraging behavioral changes and technical fixes like Transfer Learning to reduce training costs and ensure the optimization "cure" isn't worse than the power-draw "disease."

Quantitative Analysis and Results

The simulation of the O-RAN-based Cell-Free Massive MIMO system yielded significant quantitative findings regarding power efficiency. By dynamically shutting down unused GPPs and optical line cards based on user density and spectral efficiency (SE) requirements, the system achieved up to 190W of power savings at high SE levels.

Total power consumption comparison with and without O-RAN optimization across varying user densities.

As shown in the results, the scalability of O-RAN optimization is evident: as the number of users (UEs) increases, the gap between optimized and non-optimized configurations widens, reaching savings of up to 220W for 14 UEs.

Challenges and Future Directions

Despite its promise, O-RAN faces several hurdles:

  • Interoperability: Full multi-vendor interoperability is not yet a reality, and integration can currently introduce new inefficiencies.
  • Standardized Metrics: There is a lack of benchmarks to measure "social sustainability" or "energy sufficiency" for network operators.
  • AI Training Costs: Rigorous, end-to-end quantification of the energy cost of AI models is required to ensure net-positive environmental impact.

Future work must focus on bridging these gaps to ensure 6G delivers on its green promises.

Summary updated

BibTeX
@article{ahmadi2025towards,
  author = {Hamed Ahmadi and Mostafa Rahmani Ghourtani and Swarna Bindu Chetty and Eirini Eleni Tsiropoulou and Huseyin Arslan and Mérouane Debbah and Tony Q. S. Quek},
  title = {{Towards Sustainability in 6G and beyond: Challenges and Opportunities of Open RAN}},
  journal = {IEEE Commun. Stand. Mag.},
  year = {2025},
  doi = {10.1109/mcomstd.2025.3575488},
}
Figure