Systematic Literature Review: Energy Metrics, Trade-Offs, and Best Practices in Green IT and Green Software

The rapid growth of information technology, including cloud computing, the Internet of Things, and artificial intelligence, is driving increased energy consumption and carbon emissions in the technology sector. A report by the International Energy Agency (IEA) projects that global data center electricity consumption will more than double, reaching approximately 945 TWh by 2030, or about 3% of global electricity consumption. A Deloitte study even estimates that data centers could contribute around 4% by 2030, primarily driven by the growing use of artificial intelligence. Beside physical infrastructure, software also significantly contributes to energy consumption, necessitating Green IT and Green Software approaches to reduce environmental impact. This research aims to identify and classify energy metrics used in the literature, analyze the trade-offs between energy consumption and system performance, and summarize relevant best practices for application in industrial settings. This research uses the Systematic Literature Review (SLR) method, referring to the Kitchenham protocol. The search was conducted on the IEEE Xplore, ACM Digital Library, ScienceDirect, SpringerLink, and arXiv databases for articles published between 2021 and 2025 that included empirical data or conceptual models related to the energy efficiency of information systems. From the screening process, a number of articles were obtained and analyzed thematically. The research results show that energy metrics vary from hardware-based measurements (power meter) to estimation models (GSMM, GSMP). A trade-off was also found between energy savings and performance, necessitating a multi-objective optimization approach. This research contributes by providing a framework for classifying energy metrics, visualizing measurement method trends, and offering methodological recommendations that can be integrated into industrial practice. The implications of this research are to support digital sustainability efforts while simultaneously reducing the organization's operational costs.