Energy consumption in next-generation mobile networks

下一代移动网络的能源消耗

• 批准号: 2603428
• 金额: --
• 依托单位: Queen Mary University of London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2021
• 资助国家: 英国
• 起止时间: 2021 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2603428
• 关键词: Energy; consumption; next; generation; mobile

Extract taken from the Year 2 Progression Report:The rapid evolution of the telecommunications landscape and the undeniable environmental implications of its energy consumption necessitate a thorough exploration and understanding of its operational characteristics. Beginning with the current state of affairs, this project has achieved a primary goal to understand the energy consumption in current deployments. Such an understanding serves as the foundation upon which we can build and model future strategies, allowing us to gauge and measure the shifts in energy dynamics effectively.The past 12 months have yielded productive strides towards the successful development of a pre-existing tool that can model energy consumption for 5G networks, bycollaborating with telecommunications giant BT (owner of UK based mobile network 'EE'). Representing the current pinnacle of mobile communication, 5G will act as ourbaseline for comparison with potential future technologies. With the present toolset extended for energy modelling capability, we aim to explore the frontier of what lies ahead. This entails analysing the potential energy consumption in hypothetical deployments, especially those centred around the concepts central to the next generation of mobile network architectures, like Open RAN and AI-native networks. By extrapolating current 5G energy requirements towards these emergent architectures, we aim to derive insights into the scalability and sustainability of upcoming technologies. Further development will be pivotal in dimensioning energy demands and predicting future needs under varying conditions, backed by trusted real-world applicability from BT guidance. However, understanding and prediction are just parts of the solution. The real crux lies in actionable change. With the integration of AI and ML technologies, we plan to formulate strategies that can optimise energy consumption. These strategies are underpinned by a joint goal: achieving statistical significance in energy reduction and7 maximising key performance indicators for 5G and future networks. The application of AI in the RAN Intelligence Controllers (RIC) will allow for near-real-time optimisations, merging efficiency with performance.Lastly, while modelling and strategies form the theoretical basis, tangible validation in a real-world setting is essential. Therefore, our final aim is to validate the developed models and strategies against energy consumption metrics within a RAN testbed. This hands-on validation will serve as the litmus test for our proposed solutions' efficacy, robustness, and practicality.The overarching goal is to significantly reduce net energy consumption within the radio access network, especially in the context of 5G and emerging network technologies. Each aim serves as a stepping stone towards this innovation by understanding the energy demands of current deployments, devising a predictive tool for energy modelling, analysing the potential consumption of future network architectures, and using AI/ML strategies to streamline consumption. These endeavours will culminate with practical application and validation within an operational RAN framework. Ultimately, this project aims to pioneer a paradigm shift in energy efficiency, paving the way for sustainable practices in next-generation mobile networks.

摘自第二年进展报告：电信领域的快速发展及其能源消耗对环境的不可否认的影响，需要对其运营特征进行彻底的探索和理解。从目前的情况开始，该项目已经实现了了解当前部署中的能源消耗的主要目标。这种理解是我们构建和模拟未来战略的基础，使我们能够有效地衡量和衡量能源动态的变化。过去12个月，我们与电信巨头英国电信（英国移动的网络“EE”的所有者）合作，成功开发了一种既有工具，可以模拟5G网络的能源消耗。5G代表了当前移动的通信的顶峰，将作为我们与潜在的未来技术进行比较的基准。随着目前的工具集扩展到能源建模能力，我们的目标是探索未来的前沿。这需要分析假设部署中的潜在能耗，特别是那些围绕下一代移动的网络架构核心概念的部署，如Open RAN和AI原生网络。通过将当前的5G能源需求外推到这些新兴架构，我们的目标是深入了解即将到来的技术的可扩展性和可持续性。进一步的开发将是确定能源需求和预测不同条件下未来需求的关键，并得到BT指导的可靠现实适用性的支持。然而，理解和预测只是解决方案的一部分。真实的关键在于可操作的改变。随着人工智能和机器学习技术的整合，我们计划制定可以优化能源消耗的战略。这些战略都以一个共同的目标为基础：在节能方面实现统计意义，并最大限度地提高5G和未来网络的关键性能指标。人工智能在RAN智能控制器（RIC）中的应用将允许接近实时的优化，将效率与性能相结合。最后，虽然建模和策略构成了理论基础，但在现实世界中的有形验证至关重要。因此，我们的最终目标是在RAN测试平台内验证开发的模型和策略。该实践验证将作为我们提出的解决方案的有效性、稳健性和实用性的试金石，其总体目标是显著降低无线接入网络的净能耗，尤其是在5G和新兴网络技术的背景下。每个目标都是实现这一创新的垫脚石，通过了解当前部署的能源需求，设计能源建模的预测工具，分析未来网络架构的潜在消耗，并使用AI/ML策略来简化消耗。这些努力最终将在一个可操作的RAN框架内得到实际应用和验证。最终，该项目旨在开创能源效率的范式转变，为下一代移动的网络的可持续实践铺平道路。