Petabyte Storage Services For Data-Driven Science

科学价值首席研究人员(PI)将创建一种将超大规模在线磁盘存储与高性能计算紧密结合的计算系统。这一基础设施将由几个研究项目共享，每个研究项目在处理原始数据时都临时需要大量的在线存储，但在处理处理的数据时减少了存储需求。计算机图形学：PI将对物体的光散射特性进行精确测量。这将是这类研究中最准确的一次，并将导致新一代多维、多对象图形绘制技术的出现。万维网：私人投资机构将进行研究，衡量和量化万维网的发展。到目前为止，网络的巨大规模限制了对网络的报道研究只能进行一次。新的大规模存储设施将使PI能够分析多个独立的Web快照，为Web如何随时间演变开发精确的模型。天文学：PI将存储和分析来自新升级的阿雷西博射电望远镜的数据，从而发现新的脉冲星，包括具有毫秒自转周期的脉冲星，以及与其他中子星或黑洞处于双轨道上的脉冲星。脉冲星数据将为研究核物质的状态方程、引力物理、引力波、恒星演化、相对论等离子体物理以及银河系中的磁化、电离气体提供机会。存储系统将托管在康奈尔理论中心。所有项目的分析数据产品将通过互联网提供。原始数据将通过访问康奈尔理论中心或通过在磁盘驱动器上运送数据集的方式与全美的研究人员共享。私人投资促进机构制定了具体计划，将研究与各级教育活动相结合，并进行一般公众宣传。居里学院：PI将为康奈尔·居里学院开发新项目，这是一个为高中女孩提供的为期一周的暑期工程沉浸体验。第一个居里项目将使用从该设施获得的数据来说明光学和计算机图形学的基本原理。本科生教育：这项工作将在三个不同的方面影响本科生教育：(1)几个本科课程正在被调整，以利用将通过该设施获得的网络和天文数据；(2)康奈尔大学的本科生将被邀请参加与该设施相关的监督研究；以及(3)PI将增加他们与马萨诸塞州的本科生女子学院史密斯学院的互动。研究生教育：该设施自然将为博士水平的研究提供机会。此外，几个研究生课程正在计划处理正在进行的关于该设施的研究和数据密集型计算的一般领域。公共外展：公共宣传研究所的研究将在科学中心展出，这是一个由康奈尔理论中心运营的虚拟科学博物馆，可在www.SciCentr.org的网站上访问。科学中心于2004年5月通过科学和技术中心协会(ASTC)获得博物馆认证。PIS将为与新设施相关的每个研究项目开发一个科学中心展览。

Scientific Merit The principal investigators (PIs) will create a computing system that tightly couples extremely large-scale online disk storage with high-performance computing. This infrastructure will be shared among several research projects, each of which has a temporary need for large amounts of online storage when processing raw data, but has reduced storage requirements when operating on processed data. Computer Graphics: The PIs will perform precise measurement of the light scattering properties of objects. This will be the most accurate study of this kind ever undertaken, and will lead to a new generation of multidimensional, multi-object graphical rendering techniques. The World Wide Web: The PIs will perform studies measuring and quantifying the evolution of the World Wide Web. Until now, the sheer scale of the Web constrained reported studies of the Web to be performed exactly once. The new large-scale storage facility will enable the PIs to analyze multiple independent snapshots of the Web, to develop precise models for how the Web evolves over time. Astronomy: The PIs will store and analyze data from the newly upgraded Arecibo radio telescope, leading to discovery of new pulsars, including pulsars with millisecond spin periods and those in binary orbits with other neutron stars or black holes. The pulsar data will provide opportunities for research on the equation of state of nuclear matter, gravitation physics, gravitational waves, stellar evolution, relativistic plasma physics, and the magnetized, ionized gas in the Milky Way. The storage system will be hosted at the Cornell Theory Center. Analyzed data products from all projects will be made available over the Internet. Raw data will be shared with researchers throughout the U.S., either through visits to the Cornell Theory Center or by shipping data sets on disk drives.Broader Impact. The PIs have concrete plans for coupling the research with educational activities at all levels, as well as for general public outreach. The CURIE Academy: The PIs will develop novel projects for the Cornell CURIE Academy, a one-week summer engineering immersion experience for high school girls. The first CURIE project will use data obtained with the facility to illustrate fundamental principles of optics and computer graphics.Undergraduate Education: The work will impact undergraduate education in three different ways: (1) Several undergraduate courses are being adapted to make use of Web and astronomical data that will be available through the facility; (2) Cornell undergraduates will be invited to participate in supervised research related to the facility; and (3) The PIs will increase their interaction with Smith College, an undergraduate women's college in Massachusetts.Graduate Education: The facility will naturally provide opportunities for Ph.D. level research. In addition, several graduate courses are planning to address both the research being done on the facility and the general area of data-intensive computing. Public Outreach: The PIs research will be displayed in SciCentr, a virtual science museum run by the Cornell Theory Center accessible on the Web at www.SciCentr.org. SciCentr received museum accreditation through the Association of Science and Technology Centers (ASTC) in May 2004. The PIs will develop a SciCentr exhibition for each research project associated with the new facility.