DESC: Type I: Data-driven system-design for sustainable long-lasting distributed infrastructures

DESC：类型 I：数据驱动的系统设计，实现可持续的持久分布式基础设施

• 批准号: 2324873
• 负责人: Stephen Lee
• 金额: $ 52.3万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-10-01 至 2026-09-30
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2324873&HistoricalAwards=false
• 关键词: DESC; Type; Data; driven; system

To mitigate the effects of climate change, it is crucial to transition to a carbon-free economy where existing infrastructures no longer rely on carbon-based electricity. However, the rapid growth in computing demand has resulted in a significant increase in overall energy consumption, which has a detrimental impact on carbon emissions. While current research has primarily focused on reducing energy consumption, it is important to recognize that improving energy efficiency alone does not necessarily lead to reducing carbon emissions. Highly energy-efficient systems may still produce substantial carbon emissions. To effectively reduce emissions from computing systems, it is essential to consider the entire lifecycle of these systems. This project aims to develop novel lifespan-aware operation and emission-reduction strategies for distributed computing infrastructures that enable these systems to consider lifecycle emissions, promoting the longevity of computing systems and facilitating emission reduction. The project's broader impacts include developing prototypes and educational resources for K-12 and undergraduate levels. Designing distributed computing systems requires rethinking how systems can operate on clean energy sources and maintain longevity despite demand and supply uncertainties. This project addresses this challenge by designing a lifetime management system that incorporates principles of safe computing to ensure system operation and prevent degradation or failures. It focuses on developing emission-reduction strategies using novel models and policies integrated into container-based orchestration platforms. This enables applications to make sound decisions with regard to emissions and regulate their services based on the implications of these decisions. Additionally, this project also focuses on minimizing embodied emissions by extending the system's lifespan and proposing novel mechanisms and policies that adjust application resources to ensure systems operate at safe levels. Finally, techniques are developed to jointly reduce both energy use and embodied emissions, leveraging spatial and temporal variations in emissions and workload. Together, these mechanisms and policies will provide a new technical foundation for emission-aware applications.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.

为了减轻气候变化的影响，向无碳经济转型至关重要，让现有基础设施不再依赖碳电力。然而，计算需求的快速增长导致总体能源消耗大幅增加，这对碳排放产生不利影响。虽然当前的研究主要集中在减少能源消耗上，但重要的是要认识到，仅提高能源效率并不一定会减少碳排放。高能效系统仍可能产生大量碳排放。为了有效减少计算系统的排放，必须考虑这些系统的整个生命周期。该项目旨在为分布式计算基础设施开发新颖的生命周期感知操作和减排策略，使这些系统能够考虑生命周期排放，提高计算系统的使用寿命并促进减排。该项目的更广泛影响包括开发 K-12 和本科阶段的原型和教育资源。设计分布式计算系统需要重新思考系统如何在清洁能源上运行并在需求和供应不确定的情况下保持较长的使用寿命。该项目通过设计一个生命周期管理系统来解决这一挑战，该系统结合了安全计算原则，以确保系统运行并防止性能下降或故障。它专注于使用集成到基于容器的编排平台中的新颖模型和策略来制定减排策略。这使得应用程序能够就排放做出合理的决策，并根据这些决策的影响来监管其服务。此外，该项目还致力于通过延长系统的使用寿命并提出调整应用资源以确保系统在安全水平上运行的新机制和政策来最大限度地减少隐含排放。最后，开发技术来共同减少能源使用和隐含排放，利用排放和工作负载的空间和时间变化。这些机制和政策将为排放感知应用提供新的技术基础。该奖项反映了 NSF 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。