Analysis of Simulators and Adaptation Aiming at the Reduction of CO2 Emissions in Data Centers

• Grant number: 24/23163-7
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: April 01, 2025
• End date: May 31, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Daniel de Angelis Cordeiro
• Grantee: Gabriella Silva Saraiva
• Supervisor: Danilo Carastan dos Santos
• Host Institution: Escola de Artes, Ciências e Humanidades (EACH). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: Laboratoire D'Informatique De Grenoble, France
• Associated to the scholarship: 24/09487-4 - Analysis of simulators and adaptation aiming to reduce CO2 emissions in data centers, BP.IC

Abstract

Human activities increasingly depend on computer programs running on cloud computing platforms. To cope with the growing demand, the data centers used in such platforms have significantly increased their processing capacities. This increase led us to a scenario where 2% of the global energy demand in 2022 and approximately 2% of greenhouse gas emissions were attributed to these data centers. Several recent studies propose energy-aware computational resource management algorithms to reduce these impacts. However, beyond reducing energy consumption, it is essential for these algorithms to consider the carbon emissions associated with the energy mix powering these platforms, given that CO2 emissions vary significantly between different types of energy sources. This project proposes the incorporation of carbon emission calculations into simulations performed by the Batsim simulator, aiming to provide a more comprehensive analysis of the environmental impact caused by different task and resource management methods in data centers. The proposal involves the development of a plugin in SimGrid, a simulation framework on which Batsim is based, to calculate CO2 emissions based on energy consumption and the geographic location of the simulated platform, considering the composition of the local energy matrix. Subsequently, the plugin will be integrated into Batsim, ensuring compatibility with the simulations already conducted on the simulator. Finally, the feasibility of including complementary energy sources to the local power grid, such as batteries and solar panels, will be studied to enhance the carbon emissions model associated with the simulations. The project will be developed at the Laboratoire d'Informatique de Grenoble (LIG), in France, under the co-supervision of Prof. Dr. Danilo Carastan-Santos.