Abstract:Energy efficiency of data centers (DCs) is of great concern due to their large amount of energy consumption and the foreseeable growth in the demand of digital services in the future. The past decade witnessed improvements of the energy efficiency of DCs from an extensive margin—a shift from small to large, more efficient DCs. Improvements from the intensive margin, that is, from more efficient operation, would be critical in limiting the energy consumption and environmental impact of DCs in the upcoming period. Machine learning algorithms have advantages in optimizing the operation of DCs to improve energy efficiency as they have shown the potential of discovering control strategies not found by traditional method, and producing working condition‐dependent control strategies. This study proposes ready‐to‐use machine learning methods with practical details to decrease the most commonly used energy efficiency metric—power usage effectiveness in DCs. We achieved an accurate prediction by properly selecting the features used in the proposed prediction models established by neural network, light gradient boosting machine, recurrent neural network, and random forests. The proposed approaches are implemented in one of the largest hyperscale DCs in China—Tencent Tianjin DC, to optimize the set points of controllable variables in the cooling system and to detect and adjust the unreasonable working conditions in the modular data centers. The lower bound of PUE reduction was 0.005 with the proposed approaches, leading to about 1500 MWh (0.24% of the total designed electricity consumption of this DC) of energy saved per year in this hyperscale DC. The proposed methods have the potential to be transferred to DCs of similar scale, and the framework of our work could serve as a guide for machine learning‐based optimization of environmental indicators in other complex product/service systems.

Rethinking Data Center Networks: Machine Learning Enables Network Intelligence

Machine Learning Empowered Intelligent Data Center Networking: A Survey

Towards an energy-efficient Data Center Network based on deep reinforcement learning

Machine Learning for Networking: Workflow, Advances and Opportunities

A Survey on Machine Learning for Geo-Distributed Cloud Data Center Management

Bridging machine learning and computer network research: a survey

Rethinking the Data Center Networking: Architecture, Network Protocols, and Resource Sharing

Guest Editorial: Special section on Networks, Systems, and Services Operations and Management Through Intelligence

Enabling Technologies for Future Data Center Networking: a Primer

Machine Learning Empowered Content Delivery: Status, Challenges, and Opportunities

SmartFCT: Improving power-efficiency for data center networks with deep reinforcement learning

A Survey of Machine Learning for Network-on-Chips

Machine Learning Assisted Optical Network Resource Scheduling in Data Center Networks

The Need for Advanced Intelligence in NFV Management and Orchestration

Increasing the Energy Efficiency of a Data Center Based on Machine Learning

Fundamentals of Machine Learning in Data Center Networks

A Reinforcement Learning Based Data Storage and Traffic Management in Information-Centric Data Center Networks

Towards Intelligent Network Management: Leveraging AI for Network Service Detection

Generative AI in Data Center Networking: Fundamentals, Perspectives, and Case Study

Traffic Management for Distributed Machine Learning in RDMA-enabled Data Center Networks.

Wireless Network Intelligence at the Edge