DESC: Type I: Minimizing Carbon Footprint by Co-designing Data Centers with Sustainable Power Grids

数据中心目前贡献了全球温室气体排放的很大比例，估计与航空业持平。许多预测认为，由于预期的计算需求增加，到2030年，这一比例将大幅上升。环境可持续性努力的一个主要驱动力是转向可再生能源，如太阳能和风能。然而，可再生能源是间歇性的，变化无常，这构成了一个重大挑战，因为电网必须在供应(来自能源生产)和需求(来自能源消费者)之间保持平衡。该项目引入并研究了一种新的碳指标--“外源碳”，即数据中心通过提供平衡电网的服务来减少电网碳足迹的能力，同时提高可再生能源的利用率/发电量。该项目的目标是(1)估计细粒度时空碳强度并量化外源碳；(2)开发可功率延展的数据中心运行时技术，以最小化可持续电网的碳排放并最大化外源碳的效益；(3)开发数据中心设计时技术，以指导数据中心的规划和布置，以及联合设计运行时优化，以最小化生命周期的碳足迹。该项目对工业界和科学界具有重要的更广泛的影响。开发的运行时框架、管理政策和指标/工具将帮助各种规模的电网运营商和数据中心实现环境可持续性的改善，如减少碳足迹、降低能源消耗和提高电网稳定性，以实现更大的可再生能源普及率。该项目还包括一个重要的教育部分。调查人员将把研究成果和工具整合到本科和研究生课程中，并在加州大学河滨分校招收代表不足的少数族裔和本科生，加州大学河滨分校是一家经认可的拉美裔服务机构(OPEID-00131600)，以帮助培训学生为未来计算系统的环境可持续性设计基础。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Data centers currently contribute a significant percentage of the global greenhouse gas emissions, estimated to be on par with the airline industry. Many projections have this percentage increasing dramatically by 2030 due to an expected increased computational demand. A major driver of environmental sustainability efforts is the shift towards renewable energy sources, such as solar and wind energy. However, renewable energy sources are intermittent and extremely variable which poses a major challenge as power grids must maintain a balance between supply (from energy generation) and demand (from energy consumers). This project introduces and investigates a novel carbon metric, “Exogenous Carbon”, the capacity for data centers to reduce carbon footprint of the power grid by providing services to balance the power grid while increasing renewable energy utilization/generation. This project aims to (1) estimate fine-grain spatial-temporal carbon intensities and quantify exogenous carbon, (2) develop power-malleable data center run-time techniques to minimize carbon emissions from the sustainable power grid and maximize the benefit of exogenous carbon, and (3) develop data center design-time techniques to guide data center planning and placement, as well as co-design run-time optimization to minimize life-time carbon footprint.This project has a significant broader impact on industry and the scientific community. The developed runtime frameworks, management policies, and metrics/tools will help power grid operators and data centers of all sizes to realize environmental sustainability improvements, such as reducing carbon footprint, reducing energy consumption, and improving power grid stability to enable greater renewable energy penetration. This project also includes a significant educational component. The investigators will integrate the research results and tools into both undergraduate and graduate curricula, as well as recruit under-represented minorities and undergraduate students at the University of California, Riverside (UCR), an accredited Hispanic Serving Institution (OPEID-00131600), to help train students in the fundamentals of designing for environmental sustainability for future computing systems.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.