Energy-Efficient Clouds: Leveraging Performance Metrics for Sustainable Data Center Operations

Sustainable data center operations are challenged by the rising energy usage and CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions resulting from the fast expansion of cloud computing. Reducing environmental effects and running expenses depends on the best use of resources since the demand for cloud services is growing. Many times, the dynamic and varied character of workloads is disregarded by current approaches, therefore reducing their effectiveness. This work suggests a machine learning (ML) method to improve energy economy and reduce CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions in clouds. Using important performance criteria such as CPU usage, memory usage, and execution time, the framework uses complex models, including Gradient Boosting (GB) and Random Forest (RF), to precisely forecast energy consumption and execution duration. Experimental data show that GB offers better predictive performance than baseline approaches, lowering Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). Moreover, exploratory study clarifies the need for job prioritizing and workload balance for reaching energy-efficient operations. This design offers a data-driven means of best resource use, less environmental effect, and encouragement of sustainable cloud computing techniques.