MRI: Acquisition of an Adaptive Data Cluster for Data-intensive Applications in Science and Engineering

MRI：获取自适应数据集群以用于科学和工程中的数据密集型应用

• 批准号: 1429316
• 负责人: Daniel Andresen
• 金额: $ 30万
• 依托单位: Kansas State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2014
• 资助国家: 美国
• 起止时间: 2014-09-01 至 2017-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1429316&HistoricalAwards=false
• 关键词: MRI; Acquisition; Adaptive; Data; Cluster

This project, acquiring an adaptive multi-petabyte scalable storage cluster for high-end applications, aims to service multiple research areas such as: genomics, bioinformatics, computer security, digital ethnography, environmental modeling, and computer science. The flexible storage consists of an integrated multi-petabyte disk system with integrated compute capabilities and a scalable tape archive system for data and storage intensive applications. The system allows I/O-limited calculations to be performed directly on the storage nodes (unlike the traditional storage clusters) and can also act as a distributed file system with massive bandwidth to allow CPU-limited calculations to benefit from existing cluster computational resources (unlike typical Hadoop clusters). The instrument enables - Looking deeper in modeling genomes, hyper-extractive economies and phenotyping- Bringing greater data capacity tied to computational resourcing,- Attacking grand challenges (e.g., forecasting responses of ecological systems to natural and anthropogenic global and regional change).- Developing new algorithms in high-throughput phenotyping, computer security, and genome annotation (hence enabling a new science with the hybrid platform).Thus, the storage cluster constitutes a seminal component for a critical need, a campus-based facility for data immersive computing, not only at the institution, but for the entire state, since the institution currently does not have a central facility with capability for 'big data' high-end computing. As new algorithms are developed for better modeling cyber interactions that can lead to increase the financial and network infrastructure's ability, resiliency, and resistance, multidisciplinary impacts on cybersecurity research are likely to be felt. It can also contribute to train a new generation of researchers in tools and techniques for data-intensive computing and ease their migration to XSEDE when their research needs exceed the local resources. It can contribute to protect the environment when developing better models of the interaction of water, ecology, and economic factors. Moreover, it can enhance and integrate educational efforts at the K-12, undergraduate, and graduate levels in bioinformatics (e.g., preparation of educational materials, impacting the K-12 and STEM education such as 'It's a BLAST' and GROW workshops for female high school students). Ultimately, it allows access to community colleges and non-PhD granting institutions to extend big data through and EPSCoR state, enabling many researchers and educators.

该项目旨在为高端应用程序获得自适应的多PB可扩展存储集群，旨在服务于多个研究领域，如：基因组学，生物信息学，计算机安全，数字人种学，环境建模和计算机科学。灵活的存储由一个集成的多PB磁盘系统和一个可扩展的磁带存档系统组成，该系统具有集成的计算能力，适用于数据和存储密集型应用程序。该系统允许直接在存储节点上执行I/O受限的计算（与传统存储集群不同），还可以充当具有大带宽的分布式文件系统，以允许CPU受限的计算从现有集群计算资源中受益（与典型的Hadoop集群不同）。该工具能够-更深入地研究基因组建模、超提取经济和表型分析-带来与计算资源相关的更大数据容量，-应对重大挑战（例如，预测生态系统对自然和人为的全球和区域变化的反应）。开发高通量表型分析、计算机安全和基因组注释的新算法因此，存储集群构成了一个关键需求的开创性组件，一个基于校园的数据沉浸式计算设施，不仅在机构，而且在整个国家，因为该机构目前没有具有“大数据”高端计算能力的中央设施。随着新算法的开发，以更好地模拟网络交互，从而提高金融和网络基础设施的能力，弹性和抵抗力，可能会感受到对网络安全研究的多学科影响。它还有助于培训新一代研究人员掌握数据密集型计算的工具和技术，并在他们的研究需求超过当地资源时方便他们迁移到XSEDE。它可以有助于保护环境，开发更好的模型，水，生态和经济因素的相互作用。此外，它可以加强和整合K-12，本科和研究生水平的生物信息学教育工作（例如，编写教育材料，影响K-12和STEM教育，如“这是一个爆炸”和女高中生成长讲习班）。 最终，它允许访问社区学院和非博士授予机构，以通过EPSCoR状态扩展大数据，使许多研究人员和教育工作者能够。