SCI: TeraGrid Resource Partners

可扩展太尺度设施（ETF）正在开发和部署美国最大的统一计算基础设施。ETF的目标是为科学家提供尽可能强大的集成计算资源，虽然ETF资源在地理上分散，但它们代表了九个不同站点的集成能力。ETF集成的资源包括计算资源、存储、数据集合、仪器和可视化工具。ETF的用户将体验ETF作为一个单一的实用程序，不仅可以从任何地方访问，而且还可以从任何地方运行作业。ETF将努力建立一个共同的框架，以扩大和稳健可靠的运作。将邀请其他网站加入ETF，并将通过拥抱共同的软件基础设施，使ETF不仅仅是一个松散耦合的系统的HPC网站。公共基础设施是ETF的核心，它实现了网格概念，具有单一登录、单一分配货币、单一编程环境、单一服务台、单一排队系统等特性，但资源异构且分散。普渡大学作为资源提供商将通过其IBM SP和Linux集群向ETF提供100万个SU。 预计这一承诺每年将增加10%。普渡大学还将提供和维护遥感、气候建模和生物信息学方面的几个数据集。普渡大学还将与其他RP网站合作，开发一个数据收集框架，以促进元数据的发布和搜索以及大型数据集的后处理。最后，普渡大学将通过科学网关吸引科学用户，特别是通过NSF计算纳米技术网络（NCN）的计算纳米技术社区和通过普渡陆地观测站（PTO）的遥感社区。 普渡大学与两个代表主要国家和国际研究社区的科学合作伙伴一起，将促进TeraGrid资源的使用，以探索和扩展关键科学学科的前沿。我们的目标和我们成功的最终衡量标准将是建立科学网关，供这些学科的数百名科学家使用

The Extensible Terascale Facility (ETF) is developing and deploying the nations largest unified computational infrastructure. The goal of the ETF is to provide scientists with the most powerful, integrated computational resource possible, While ETF resources are geographically dispersed, they represent the integrated capability of nine different sites. The resources integrated by ETF include compute resources, storage, data collections, instruments, and visualization tools. Users of the ETF will experience the ETF as a single utility that is not only accessible from anywhere, but will also run jobs from anywhere. The ETF will work towards building a common framework for expansion and robust reliable operation. Additional sites will be invited to join the ETF and will do so by embracing the common software infrastructure that makes ETF more than a loosely coupled system of HPC site. The common infrastructure is at the heart of the ETF and brings the grid concept to fruition with features such as single login, single allocation currency, single programming environment, single helpdesk, single queuing system but heterogeneous and dispersed resources. Purdue University as a Resource Provider will make available to the ETF, 1,000,000 SUs through its IBM SP and Linux clusters. This commitment is expected to increase by 10% each year. Purdue will also make available and maintain several dataset collections in remote sensing, climate modeling, and bioinformatics. Purdue will also work with other RP sites to develop a data collection framework that will facilitate metadata publishing and searching as well as post-processing of large datasets. Finally, Purdue will engage scientific users through science gateways, especially the computational nanotechnology community through the NSF Network for Computational Nanotechnology (NCN) and the remote sensing community through the Purdue Terrestrial Observatory (PTO). With two science partners representing major national and international research communities, Purdue will stimulate the use of TeraGrid resources to explore and extend the frontiers of key scientific disciplines. Our goal and the ultimate measure of our success will be establishing science gateways used by hundreds of scientists in each of these disciplines