Collaborative Research: Framework: Data: Toward Exascale Community Ocean Circulation Modeling

该项目设计并实现了一个用于处理PB级数据集的软件框架；重点是全球海洋环流。由三所大学(约翰·霍普金斯大学、麻省理工学院和哥伦比亚大学)组成的团队建立了一个统一的数据系统，该系统能够以1公里的水平分辨率提供全球海洋环流模式输出。该产品将托管在一个开放的门户中，为社区提供可扩展的软件工具，以实现对数据集的分析。研究小组将利用这些数据来回答有关海洋中的混合和消散过程的具体问题。这项工作的目标是创建和演示一个完整的、可复制的网络基础设施，用于共享和分析大规模模拟。重点是高分辨率海洋环流建模，以及能够实现高效存储的软件工具。讨论了海洋和气候动力学研究的两个主要挑战：处理来自高分辨率模拟的大数据集，以及理解小尺度海洋过程在大尺度海洋/气候系统中的作用。解决第一个挑战将极大地促进目前和今后对海洋/大气/气候系统的研究；解决第二个挑战将深刻地增进对海洋/气候动态的了解。该项目建立了一个统一的数据系统，包括高分辨率的全球海洋环流模拟、用于数据共享的千万亿级门户和用于交互分析的可扩展软件工具。该项目的软件框架预计将为用户处理千万亿级到亿级级的数据集。该项目利用了几个已有的能力：东格陵兰大陆坡溢油急流的JHU区域数值模式、来自Pangeo项目的软件、本服务器数据密集型软件基础设施以及从东北存储交换多PB区域数据存储中吸取的经验教训。更广泛的目标是地球科学和其他学科的下一代模拟软件。该奖项由美国国家科学基金会高级网络基础设施办公室颁发，由海洋科学部和国家科学基金会地球科学局内的综合与合作教育与研究计划共同支持。该奖项反映了美国国家科学基金会的法定使命，通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project designs and implements a software framework for handling petabyte-scale datasets; the focus is on global ocean circulation. A team of three universities (Johns Hopkins University, MIT, and Columbia University) builds a unified data system that is capable of delivering global ocean circulation model output at 1 km horizontal resolution. The product will be hosted in an open portal, providing the community with scalable software tools to enable analysis of the dataset. The team will use this data to answer specific questions about mixing and dissipation processes in the ocean. The goal of this effort is the creation and demonstration of a complete and replicable cyberinfrastructure for sharing and analysis of massive simulations. The focus is on high resolution ocean circulation modeling, with software tools that will enable efficient storage. Two major challenges to the study of ocean and climate dynamics are addressed: handling large datasets from high-resolution simulations, and understanding the role of small-scale ocean processes in large-scale ocean/climate systems. Resolving the first challenge would significantly facilitate ongoing and future studies of the ocean/atmosphere/climate system; addressing the second challenge would profoundly improve understanding of ocean/climate dynamics. The project builds a unified data system consisting of high-resolution global ocean circulation simulations, a petascale portal for data sharing, and scalable software tools for interactive analysis. The software framework from this project is expected to handle petascale to exascale datasets for users. Several pre-existing capabilities are leveraged for this project: the JHU regional numerical model of the Spill Jet on the East Greenland continental slope, software from the Pangeo project, the SciServer data-intensive software infrastructure, and lessons learned from the North East Storage Exchange multi-petabyte regional data store. The broader target is next generation simulation software in the geosciences and other disciplines. This award by the NSF Office of Advanced Cyberinfrastructure is jointly supported by the Division of Ocean Sciences and the Integrative and Collaborative Education and Research Program within the NSF Directorate for Geosciences.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.