The National Science Foundation's (NSF) Directorate for Computer and Information Science and Engineering (CISE) advances understanding of the principles and uses of next-generation computing, communications, and information systems in service to society; supports advanced cyberinfrastructure that enables and accelerates discovery and innovation across all science and engineering disciplines; and contributes to universal, transparent, and affordable participation in an information-based society. CISE provides approximately 80% of the federal funding for fundamental computer science research at U.S. academic institutions.
Andrew A. Chien's Communications editorial "Computing's Grand Challenge for Sustainability"a (Oct. 2022) lays out a compelling and timely challenge for environmental sustainability: "The computing community should embrace a grand challenge to reduce the carbon-emissions and environmental impact of computing in absolute terms dramatically, and if possible, to zero." The role of computing and environmental sustainability, as noted in Chien's editorial, can be viewed in two different ways: "Computing in Sustainability" and "Sustainability in Computing." This Viewpoint focuses on ways the computing community can contribute broadly to environmental sustainability and identifies CISE research programs supporting these efforts.
Computing in sustainability leverages advances in computing technology to understand and analyze the climate ecosystem, build resilience to climate-driven extreme events, and mitigate and adapt to climate change. These techniques include:
While some of these capabilities already exist and have been applied to climate problems (for example, smart agriculture), more multidisciplinary research is needed to develop an integrated approach that will work at scale and have true impact on climate-driven problems. In addition, access to advanced cyberinfrastructure (for example, democratizing science through advanced cyberinfrastructure,b and National Discovery Cloud 2021c) can expand the geography of innovation by enabling researchers everywhere to engage fully in and make advances in these research areas.
Computing in sustainability leverages advances in computing technology to understand, analyze, and build resilience in the climate ecosystem.
Much of this foundational work is supported by the CISE core research programs,d as well as other NSF research programs focused on specific aspects of sustainability: for example, the National AI Research Institutese program has multiple themes that explore the role of AI in sustainability; the Cyber Physical Systemsf and Smart and Connected Communitiesg programs support exploration of more effective adaptation and mitigation of climate effects; and the CIVIC Innovation Challengeh program is focused on improving resilience to climate events driven by community priorities. Furthermore, programs such as the Cyberinfrastructure for Sustained Scientific Innovationi and Advanced Computing Systems & Services: Adapting to the Rapid Evolution of Science and Engineering Researchj support the development of sustainable research cyberinfrastructure resources crucial for enabling fundamentally new scientific and engineering advances.
Sustainability in computing, the focus of Chien's editorial, is concerned with improving the sustainability and mitigating the climate impacts of the computing ecosystem itself. Much of the research in sustainable computing in the last decade has focused on operational efficiency, but embodied energy, such as carbon dioxide and other harmful emissions associated with materials and manufacturing processes, can often be greater than the operational energy. In order to truly respond to Chien's grand challenge, one must consider broader notions of sustainability including significantly reducing greenhouse warming gas emissions (GHGs), use of volatile organic compounds (VOCs), and consumption and disposal of rare materials, heat, and wastewater; while improving the recyclability, reuse, and longevity of these systems.
Sustainability in computing is concerned with improving the sustainability and mitigating the climate impacts of the computing ecosystem.
The goal of the newly released program solicitation for the Design for Environmental Sustainability in Computingk program is to reduce the substantial environmental impact that computing poses through its entire life cycle from design and manufacturing, through deployment into operation, and finally into reuse, recycling, and disposal. As Chien points out, "This daunting challenge is rife with hard research problems" and will require transformative solutions that address sustainability at all layers of the system stack and all steps in the system life cycle from new hardware and network architectures to sustainability-aware algorithms, software and system design to sustainable management of increasingly large datasets and workloads to end-user decision making around adoption, use, repurposing, and ultimately disposal of computing systems.
CISE welcomes the upcoming Communications special section, as mentioned in Chien's editorial, which will focus on sustainability and the deep challenges computing faces, with the goal of inspiring research and new ideas. We look forward to working with the research community in moving the needle on environmental sustainability for computing and beyond.
Acknowledgments. The authors thank Margaret Martonosi, Joydip Kundu, Behrooz Shirazi, Alex Jones, and Goli Yamini for helpful comments and suggestions.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2023 ACM, Inc.
Terrific to see both this viewpoint and the NSF's initiatives in both "Computing for Sustainability" and recently "Sustainability of Computing"! These are two critical roles and challenges for computing to be a positive force for environmental sustainability.
I applaud and support the broader focus on "embodied carbon" (not embodied energy) and lifecycle impacts of electronics and computing. This broader scope is critical to progress finding ways to reduce computing's negative environmental impact in the face of exponentially increasing use of computing (eg. digitalization, AI/ML, pervasive intelligence, and move)!
Its worth observing that operational and embodied carbon are closely coupled -- TSMC's operational carbon turns into Apple's product lifecycle embodied carbon, etc. Worse, large differences in embodied carbon may be possible with "supplier selection" rather than computing technical design choices. This makes it essential to root design and metrics for research in data that is supplier neutral... a challenge for research. For example, how do we characterize a 5nm process? 3nm process? stably, in a fashion independent of vendor. Would a design company (eg Nvidia, Apple, Qualcomm) select a foundry provider based on sustainability metrics over silicon capability?
In the other direction, we can work on embodied carbon by reducing operational carbon in the supply chain (eg decarbonize Taiwan's or South Korea's power grid), or the use phase (use low-carbon power).
Note: The coupling of operational and embodied carbon is a design feature of the GHG (green house gas) reporting protocols, not a defect. The GHG protocol designers' objective is to drive overall societal reductions, for which business organization and vendor selection is a critical element.
cheers, -Andrew
Displaying 1 comment