A Dataset for Supporting Power Management Experimentations in Cloud and Edge Infrastructures

As advancements in cloud computing, edge computing, and mobile telecommunication infrastructures continue to accelerate, concerns regarding environmental sustainability have become increasingly important. In this context, optimizing the energy efficiency of computational resources, particularly CPUs, which are major contributors to system-level power consumption becomes a key area of focus. This paper introduces a new, publicly available dataset that captures the effects of CPU P-States, C-States, and frequency Governors on power consumption and system performance. The dataset comprises 5,621 unique configurations with 337,260 samples of these CPU parameters, each tested under three distinct load conditions: idle, medium, and high. These scenarios are designed to emulate real-world usage patterns observed in cloud and edge infrastructures. By enabling in-depth analysis of performance/energy trade-offs, the dataset serves as a valuable resource for researchers and practitioners aiming to develop sustainable computing solutions.