The exponential growth of cloud computing has led to a significant increase in the number and energy intensity of data centers, which now play a crucial role in providing on-demand computing resources. This review focuses on techniques such as host CPU utilization prediction, underload/overload detection, virtual machine (VM) selection, migration, and placement, employing machine learning, heuristics, metaheuristics, and statistical methods. The findings indicate that heuristic approaches have achieved energy savings ranging from 5.4% to 90% compared to existing methods. Metaheuristic techniques have demonstrated a reduction in energy consumption from 7.68% to 97%, while machine-learning methods have shown savings from 1.6% to 88.5%. Statistical methods have also contributed, reducing energy use by 5.4% to 84% when benchmarked against various approaches under diverse settings and parameters. The overarching goal of this review is to synthesize the diverse methodologies researchers have employed to enhance energy efficiency in cloud data centers, thereby contributing to more sustainable resource management. Keywords: Computing resources, Pay-as-you-go, Cloud-data, Metaheuristic, Heuristic, data-centers, Sustainability.
Salem Fadlalla Abdalhamid Mansori (Tue,) studied this question.