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.

ECDX: Energy Consumption Prediction Model Based on Distance Correlation and XGBoost for Edge Data Center

ForeXGBoost: Passenger Car Sales Prediction Based on XGBoost

Prediction Method of Energy Consumption Based on Multiple Energy-Related Features in Data Center

FS-DBoost: cross-server energy efficiency and performance prediction in cloud based on transfer regression

Building Energy Consumption Prediction: An Extreme Deep Learning Approach

GreenBytes: Intelligent Energy Estimation for Edge-Cloud

Cloud-edge collaboration based transferring prediction of building energy consumption

XGBoost energy consumption prediction based on multi-system data HVAC

An Integrated Method for High-Dimensional Imbalanced Assembly Quality Prediction Supported by Edge Computing

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

Energy-effective IoT Services in Balanced Edge-Cloud Collaboration Systems

Power Consumption Prediction with K-nearest-neighbours and XGBoost Algorithm.

Prediction of Overall Energy Consumption of Data Centers in Different Locations

An AI Model Automatic Training and Deployment Platform Based on Cloud Edge Architecture for DC Energy-Saving

5G Ultradense Cellular-Network-Based Edge Demand Response: Energy Consumption Reduction

Residential Electricity Consumption Prediction Method Based on Deep Learning and Federated Learning Under Cloud Edge Collaboration Architecture

A Collaborative Cloud-Edge Approach for Robust Edge Workload Forecasting

Learning-based power prediction for data centre operations via deep neural networks

An Effective Prediction of Resource Using Machine Learning in Edge Environments for the Smart Healthcare Industry

EsDNN: Deep Neural Network based Multivariate Workload Prediction Approach in Cloud Environment