The Next Generation Mobile Networks Alliance (NGMN) has published “Green Future Networks: A Roadmap to Energy Efficient Networks”, offering new guidance and recommendations on reducing energy consumption in mobile networks.
The roadmap aims to enhance energy-saving methods for MNOs and the wider industry by outlining 16 different energy saving techniques that are currently used or under development in the industry.
According to the publication, energy consumption can be reduced through process optimisations, engineering and operational improvements, and the deployment of recent technologies. This is the latest phase of NGMN’s Green Future Networks programme, building on the previous publications that addressed the short-term solutions that mobile network operators (MNOs) could deploy.
One new addition is on the use of Artificial Intelligence (AI) and Machine Learning (ML) to better plan and manage networks, and match planned and operational capacity to predicted traffic.
Laurent Leboucher, Member of the NGMN Alliance Board and Group CTO at Orange said, “The solutions span multiple domains: better network planning and engineering, improved network management, the application of Artificial Intelligence and Machine Learning, and the development and use of new technologies. Only by working together and collaborating within industry alliances such as NGMN can we achieve these goals.”
AI and network energy efficiency
To support network EE optimisation, NGMN said that Artificial Intelligence (AI) could be a key tool to provide Energy Consumption (EC) estimation and prediction while limiting the amount of data collected and transferred throughout the network. The report recommends standards organisations define methodologies to transfer and update AI models at the nodes where network configuration and parameters are controlled. In this context, with the increasing size of AI models, we advocate the need for solutions able to adjust the model complexity to minimise the EC that arises from: model training, transmission and execution.
NGMN also highlights that, soon, MNOs may seek to integrate novel hardware and software mechanisms to support AI-based network EE modelling and optimisation. At the software level, integrating new intelligent solutions could allow the network to reduce EC by adjusting the available network capacity to the actual traffic load at each given point in time. In indoor deployments, a new energy saving technology is proposed to manage the state of each RU all belonging to a given cell independently to dynamically switch off the Power Amplifier (PA) of any RUs that do not harbour user connection or data transmission at a given point of time. Trials have highlighted that this solution achieves an energy saving gain of 20% relative to an always-on network deployment.
Where switching off radio components is not possible, due to non-negligible load, this report shows that a RU implementing an intelligent resource allocation that decreases the transmit power by limiting the transmission spectral efficiency could lead up to 30% reduction in load-dependent EC at the RU, without impacting users’ Quality of Service (QoS). We recommend standards organisations to define methodologies to coordinate properly RUs tasked with implementing distinct and potentially competing energy saving mechanisms.
Co-ordination and sharing
Further EE gains could be realised through RAN solutions that implement different levels of coordination to achieve a more efficient usage of network resources. More specifically, this publication highlights trials related to a novel network optimisation approach that, leveraging heterogeneous QoS requirements in the service area, results in up to 18% of reduction in the average RAN EC. In addition, this report presents network level solutions where multi-carrier coordinated scheduling and spectrum sharing are respectively combined with cell discontinuous transmissions and carrier shutdown to lead up to 26%, EE gain at the RUs.
In addition to coordination mechanisms implemented through new software functionalities, MNOs are encouraged to share part of their wireless infrastructure to reduce component duplications and jointly utilise network resources through RAN sharing to limit EC and carbon emissions.
Renewable sources
As reported in the previous publication of the Green Future Networks Programme the usage of renewable energy sources is a crucial solution for reducing carbon emissions and limiting the mobile network dependency on electrical grids. This report also highlights the need for solutions to jointly dimension the power supply and network communication resources. We recommend standards organisations to enhance interworking between mobile networks and the energy suppliers to effectively reduce costs and the respective carbon footprints, while maintaining service availability.
Open RAN
Virtualised and disaggregated mobile networks speed up network deployment and
management which can improve operational efficiency. In this context, this publication overviews the current state of the O-RAN ecosystem and the actions envisaged to accelerate the deployment of energy efficient networks.
AI
The report highlights practical applications of AI and its related challenges within mobile networks. It shows that AI algorithms could help to make better energy saving decisions, such as controlling the energy saving policy thresholds, by predicting future EE and load states as well as by identifying low-EE sites.
NGMN advises that MNOs carefully assess the solutions to enhance network energy efficiency presented in this publication and analyse the prioritised list of available strategies presented in the Conclusion section.
The landscape for Energy Saving Innovation is clearly laid out – MNIS will track the industry response.
By Keith Dyer, Editor, The Mobile Network & Programme Director, Mobile Network Innovation Summit
