April 24, 2025
Abstract:
As the global demand for digital infrastructure grows, data centers face increasing pressure to minimize their environmental impact, particularly regarding energy consumption and carbon emissions in the era of AI. This talk explores approaches to enhance the sustainability of data center operations, with a focus on practical carbon footprint analysis and optimization strategies. Key topics include addressing carbon emission reporting challenges and data gaps in relation to the credibility of net-zero claims from big-tech companies and their future projections. We introduce tools designed to quickly assess both operational and embodied carbon footprints, enabling data centers to make informed decisions about sustainability. Finally, we present techniques to explore the potential for optimizing workload migrations to reduce electricity costs and carbon emissions, alongside evaluating AI workloads for demand response participation to better align energy usage with grid demands. Through these insights, this talk aims to highlight strategies and techniques for achieving greater energy efficiency and carbon reduction in data center operations, driving progress toward a more sustainable digital future.
Bio:
Dr. Can Hankendi is a post-doctoral researcher at Boston University, specializing in sustainable and carbon-aware computing, as well as energy-efficient management of compute platforms. His research focuses on optimizing data center operations for energy and carbon efficiency and exploring data center participation in demand response programs.Prior to joining Boston University as a post-doctoral researcher in 2024, Dr. Hankendi held several roles in the industry from 2013 to 2021. He served as a Power Management Architect at AMD, Power/Thermal Architect at Samsung Austin Research Center and Post-silicon Power/Thermal Optimization Engineer at Qualcomm. He was also the founder of a startup company, where he built SaaS solutions for financial institutions and retail investors. He successfully exited his startup in 2023.Dr. Hankendi holds a Ph.D. degree in Computer Engineering from Boston University (2015) and a M.Sc. degree in Electrical & Computer Engineering from University of Southern California (2010).