NSF Workshop on Sustainable Computing
AI, Water, and Biodiversity

August 20-21, 2024
Purdue University, West Lafayette, IN

August 20th, 2024

August 21st, 2024

Speaker and Talk Info

Keynote 1: Chandra Krintz (University of California, Santa Barbara)

Title: Sustainable IoT Systems for Digital Agriculture

Abstract: The Internet of Things (IoT) has great potential for extending human perception and automating intelligent actuation and control of cyber-physical systems. However, its adoption and wide spread use is nascent. A key reason for this is that IoT deployments are vastly heterogenous with widely varying resource constraints which makes them very challenging to program, deploy, secure, and maintain. Further, many of the most compelling use cases from climate resilience, disaster prediction and response, ecology and agriculture, etc. require deployments that are remote, inaccessible, hostile to electronics, that lack power infrastructure (forcing reliance on batteries), and span trust domains (edge versus cloud). Our research focuses on new approaches that simplify the development and management of IoT applications in these settings. In particular, we have designed a new programming system that makes AI/ML applications portable across the edge-cloud continuum, robust to faults and intermittent connectivity, and easier to deploy and manage. We enable this through a novel combination of a distributed serverless runtime, dataflow programming abstractions, and intelligent deployment optimizations. We design, test, and evaluate these advances using real applications for digital agriculture (as part of the UCSB SmartFarm project). We find that they simplify IoT software development and deployment while reducing energy consumption significantly.

Bio: Chandra Krintz is a Professor of Computer Science at the University of California, Santa Barbara (UCSB). She joined the UCSB faculty in 2001 after receiving her M.S. and Ph.D. degrees in Computer Science from the University of California, San Diego. Chandra has led a number of different research projects that have advanced the state-of-the-art in programming and distributed systems in ways that improve performance and energy consumption, and that ease development and deployment of software. Recently, her work has focused on the intersection of IoT, edge and cloud computing, and data analytics with applications in farming, ranching, and conservation science (cf SmartFarm and WTB). Chandra has advised over 80 undergraduate and graduate students, has published numerous research articles regarding the implementation of programming languages, participates in efforts to broaden participation in computing, and is the progenitor of the AppScale project. Chandra's efforts have been recognized with a NSF CAREER award, the CRA-W Anita Borg Early Career Award, the UCSB Academic Senate Distinguished Teaching Award, and as a UCSB Sustainability Champion.

Keynote 2: Andrew A. Chien (University of Chicago and Argonne National Laboratory)

Title: Reduce Datacenter Operational Carbon and leverage that Efficiency to reduce the Water-use and Embodied Carbon

Abstract: AI and Cloud computing is growing rapidly and projected to increase US electric power consumption by as much as 2% by 2026 (EIA). Such rapid growth is a significant challenge to power grid stability and decarbonization. With datacenter loads at 15%, even 25% in several grids, datacenter growth is browning the grid. We will discuss how computing flexibility could help the grid decarbonize (Zero-carbon cloud), and how generative AI applications such as ChatGPT can be greened. The key is cooperation between datacenter loads and power grids. Operational carbon, water-use, and embodied carbon are often viewed as disconnected sustainability challenges. They are intimately connected. We illustrate this connection, and first show how the progress in operational carbon can be used to reduce datacenter water-use and associated biodiversity loss. Second, we will show how reductions in operational carbon can be used to reduce the embodied carbon and increase the carbon-efficiency of computing. More information available at http://zccloud.cs.uchicago.edu/.

Bio: Andrew A Chien is the William Eckhardt Distinguished Service Professor of Computer Science at the University of Chicago and Senior Scientist at Argonne National Laboratory. He has led the Zero-carbon Cloud project since 2015, and is well-known for his research on datacenters, renewable energy and sustainability, cloud resource management and software, large-scale system architecture, and graph computing architecture. He is leader of the IARPA funded “UpDown System Project”, designing breakthrough scalable graph analytics systems. Chien has received numerous recognitions for his research. Dr. Chien currently serves on the NSF CISE Advisory Committee, NSF Advisory Committee on Advanced Cyberinfrastructure, and DARPA ISAT. He is a Fellow of the ACM, IEEE, and AAAS. He served as EiC of Communications of the ACM, 2017-2022, and Vice President of Research at Intel Corporation from 2005-2010. He has served on the Faculty of the University of Illinois and as SAIC Chair Professor of University of California, San Diego. He received BS, MS, and PhD degrees from the Massachusetts Institute of Technology.

Invited Talk 1: Shaolei Ren (University of California, Riverside)

Title: The Hidden Water Footprint of AI

Abstract: The growing carbon footprint of artificial intelligence (AI) models, especially large ones such as GPT and LLaMA, has been undergoing public scrutiny. However, the equally important and enormous water footprint of AI models has mostly remained under the radar. Importantly, the global AI demand is projected to account for 4.2 -- 6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4 -- 6 Denmark. This is very concerning, as freshwater scarcity has become one of the most pressing challenges shared by all of us in the wake of the rapidly growing population, depleting freshwater resources, and aging water infrastructures. To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint. In this talk, I will present a principled methodology to estimate AI’s water footprint, followed by a software-based approach to reducing the water footprint by exploiting spatial-temporal diversities of AI models' runtime water efficiency. Finally, I will discuss how to holistically and equitably minimize AI’s environmental footprint to enable truly responsible AI.

Bio: Shaolei Ren is an Associate Professor of electrical and computer engineering at the University of California, Riverside. His research broadly focuses on addressing algorithmic and systems challenges to build responsible AI, with applications to climate change mitigation, networked system designs, and socially equitable computing. His work on sustainable and equitable AI has generated significant societal impacts, including informing AI policies (e.g., adopted in AI governance guidelines of many international organizations and governments such as the United Nations, UNESCO, WHO, UN Trade and Development, OECD, Government of Canada, and Government of the Netherlands), advancing K-12 education (e.g., included in international and US high-school course materials offered by Financial Times to nearly 5,000 schools worldwide), leading to industry innovations (e.g., the industry's first real-time water footprint reporting tool for computing), and increasing public awareness of responsible AI (e.g., covered by 1,000+ major media outlets in ~100 countries, including The Associated Press, The Wall Street Journal, and Nature). He received the NSF CAREER Award in 2015 and multiple paper awards including ACM e-Energy’24 and ’16 and IEEE ICC'16. He holds a Ph.D. degree from the University of California, Los Angeles, and currently serves on the steering committee of the US Center for Strategic and International Studies (CSIS) Project for Water Security.

Invited Talk 2: Teresa Nick (Microsoft)

Title: Data Center Cooling: Environmental Life Cycle Assessment

Abstract: Optimizing data center cooling is a pivotal step towards sustainable computing and AI. The complexity of cooling optimization lies in balancing energy, water, and PFAS "forever" chemicals. Our Life Cycle Assessment (LCA) compared four cooling systems—air, two-phase immersion, one-phase immersion, and cold plates—across general purpose data centers. We included details across software, IT hardware, support equipment, and the critical environment, simulating climate and weather at two locations. Our evaluation focused on energy use, greenhouse gas emissions (GHGs), and blue water consumption. Preliminary findings pinpointed the most efficient units (virtual core or floating point operation), the significant effect of IT hardware refresh rates, and the scarcity of Environmental Product Declarations for data center components. Initial results indicate liquid cooling methods outperform air cooling, by reducing energy, emissions, and water usage.

Bio: Teresa Nick, Microsoft’s Director of Natural Systems & Sustainability, investigates datacenter environmental impacts and innovates low-energy networks. Her previous role as Director of Sustainability Information Management saw her spearheading datacenter sustainability analytics. Teresa’s academic journey includes a Neuroscience PhD from Yale, a Caltech postdoc, and academic tenure at the University of Minnesota Medical School. At Microsoft, Medtronic, and McKinsey, she’s led transformative innovation programs, promoting eco-positive business practices. Teresa’s accolades include the iM100 Award from Infrastructure Masons in 2023 and contributions to the NSF’s CIVIC Climate program. Her expertise spans IoT, wearables, and real-time algorithms, with 10+ patents, 20+ scientific papers, and 15+ digital certifications like CISSP and PMP. Her work in sustainable networks and datacenters reflects her dedication to eco-friendly innovation.

Invited Talk 3: Ben Townsend (Google)

Title: Climate Conscious Data Center Cooling and Managing the Water-Energy-Carbon Nexus

Abstract: Responsible Water Use and Google's Framework for evaluating local watershed health. Water Replenishment and our mission to replenish 120% of the water we consume by 2030.

Bio: Ben Townsend, Google global head of infrastructure and sustainability. Ben leads delivery of Google's Net Zero, Water Sustainability, and Circular Economy commitments across Google's global data center infrastructure portfolio. He received his BS and MS degrees in Civil and Environmental Engineering from Purdue University.