Grant proposal resources appearing on this page and in links embedded in its sections can be used as templates or as ancillary material for grant proposals. The documents are available in text (TXT), Microsoft Word DOC format, and PDF. Materials are grouped by major categories of services offered by Duke Research Computing and, below those sections, by NSF and NIH funding agencies, which sometimes have special requirements for proposals. Many of the documents were adapted from sections of funded proposals and drafts are dated so that you can judge their currency. Duke Research Computing staff will gladly help you tailor these documents to fit the circumstances of your proposal. Duke Research Computing staff can also be included in proposals when their talents meet needs.

Letters of support

Duke Research Computing routinely provides letters of support for NSF and NIH grant proposals in order to show institutional support for research proposals. Letters of support outline services that Duke Research Computing will offer, and staff are eager to discuss proposals and their computing, data storage, and data management requirements well before deadlines approach. Staff will also assist with detailed information for budget justifications relating to information technology.

Contact rescomputing@duke.edu for additional information about obtaining a letter of support, assistance with grant preparation and budget justifications, or to get more information on any of the items on this page.

Computing cycles

Duke Research Computing provides researchers computational resources on various platforms, each suited to fit different types of computing task. The Duke Compute Cluster (formerly called the Duke Shared Cluster Resource or “DSCR”) provides high performance/high throughput computing, and diverse offerings of “virtual machines” (VMs) provide small computers for occasional and brief use and others provide highly tailored and persistent computational set-ups for groups and for individuals. Small VMs can be procured on demand through Virtual Computing Lab and VM-Manage services, both of which have web-based provisioning tools. Larger and more specialized VMs can be ordered through the Office of Information Technology or Duke Research Computing (rescomputing@duke.edu).

Duke Compute Cluster (formats: TXT, Microsoft Word DOC, PDF) is a high-performance/high-throughput compute cluster that works well for computational tasks that can be run in parallel. The cluster is essentially a community-driven resource, with compute nodes provided by researchers and by the University. The University, through services of the Office of Information Technology, configures, updates, and maintains the nodes through their useful life. Nodes provided by researchers are available to them at “high priority”; researchers who have contributed nodes also have “low priority” access to others’ nodes, yielding to jobs with high priority. The cluster has been operating continuously since 2002. (Template updated January 2017.)

Virtual Machines (VMs) – VM-Manage, RAPID (Research Toolkits), and specialized computation (formats: TXT, Microsoft Word DOCX, PDF). VM-Manage and “RAPID” virtual machines are procurable through a web interface.  VM-Manage virtual machines can be reserved for a complete semester, with a renewal option, and their states can be saved and even transported to other environments, such as a larger virtual machine or a cloud vendor. VM-Manage machines have small RAM and CPU resources, which are sometimes constraints for research computation. Research Toolkits’ on-demand RAPID VMs are a new service was piloted in 2015, going into “production” later that year. Research Toolkits is a web-browser-based interface for Duke researchers to set up projects, establish teams of people, and provide computing and data storage resources. Provisioning of virtual machines is very flexible, with RAM sizes and CPU counts determined by faculty members of their assignees. All regular rank faculty at Duke have default RAM, storage, and CPU-core allocations that are configured through web-based tools on Research Toolkits (https://rtoolkits.web.duke.edu/, Shibboleth authentication required)

Virtual machines also can be provisioned in the Protected Data Network to allow analysis of sensitive data. Virtual machines are widely used at Duke as dedicated computational servers, providing researchers easy access to computational resources that are well fit for their specific needs. In addition, costs of virtual machines have declined significantly year-over-year. A description of virtual machine offerings is available on the OIT website [oit.duke.edu]. Information about costs and configuration options is available through Duke Research Computing — rescomputing@duke.edu. Template updated in May 2017.

Open Science Grid / Duke CI-Connect. The Open Science Grid (OSG; twiki.opensciencegrid.org) is “a national, distributed computing partnership for data-intensive research” with resources that are particularly well suited to computational jobs that have very simple systems and data dependencies — essentially, jobs that are easy to transport, run in parallel, use or produce small or modest amounts of data, and can tolerate failure or interruption. Thousands of computers in North America are part of the grid, and together they provide prodigious computational power. The Open Science Grid provided researchers nearly 800 million CPU hours in 2014 (current usage graph [display.grid.iu.edu]). Duke is part of CI-Connect [ci-connect.net] which is a means of gaining access to the Open Science Grid, and accounts are available to members of the Duke research community (http://duke.ci-connect.net). OSG is supported by the National Science Foundation and the Department of Energy’s Office of Science.

Data storage

Duke Data Commons (NIH 1S10OD018164-01) was put into operation in November 2014 as a large scale data storage resource for several of Duke’s data-producing core facilities serving life sciences research. Total capacity of the equipment is about 1.5 petabytes, and computational resources for data analysis can be attached to the storage. Although the installation is particularly targeted to support NIH-sponsored life sciences research, projects with extraordinary data storage requirements may also use the equipment. Description formats: TXT, Microsoft Word DOC, PDF. Template updated December 2015.

Duke’s Office of Information Technology offers a broad array of data storage options that are suited for a variety purposes linked to different levels of service. As is the case with costs for virtual machines, data storage costs drop significantly year-over-year, historically between 15 and 25 percent.

Data management

Protected Data Network description from the Duke University Information Technology Security Office (ITSO). Additional general information on the Protected Data Network is available on the Office of Information Technology website, and the ITSO’s website offers more detailed policies and procedures (Shibboleth authentication may be required for some documents).

Social sciences researchers can work with the Social Sciences Research Institute (SSRI; http://ssri.duke.edu/) to use the Protected Research Data Network (PRDN) that is specially outfitted for social sciences research. The Institute has professional staff who help researchers navigate special requirements for sensitive social sciences research data, whether the data are bound by government regulation or other data use agreements. The SSRI’s PRDN is a special section of the University’s Protected Data Network, with policies and technical controls tailored for social sciences research.

Data Management Plan. See the Duke University Library’s collection of information on “NSF data management plans” and “Planning for data management across the project life cycle” [library.duke.edu].  DMPTool.org provides helpful guidelines for creating “data management plans” (the “DMP” in the name of the tool). Duke University is a member of the DMPTool group, and tailored responses based on Duke’s infrastructure can be accessed on the site. Access is provided via Shibboleth authentication, and users have access to previously completed data management plans. Publicly available sample data management plans are also available on the website. The data management plan requirements for many federal agencies and private philanthropies are also available (see “DMP Requirements” on the site’s home page).

NSF proposal-styled resources

Data Management Plan. See “Data management” above.

Campus CI Plan / Campus Cyberinfrastructure Plan (formats: TXT, Microsoft Word DOC, PDF). Many NSF proposals require a “Campus Cyberinfrastructure Plan” — especially proposals in fields that involve development of computing infrastructure or areas of computer science. The document provides the agency and grant reviewers a broader perspective on the computing context in which the proposed project will take place.

Facilities statement relating to information technology (formats: TXT, Microsoft Word DOCX, PDF). The description describes physical plant relating to IT. Updated April 2017.

NIH proposal-styled resources

Data Management Plan. See “Data management” above.

Facilities statement relating to information technology (formats: TXT, Microsoft Word DOCX, PDF). The description describes physical plant relating to IT. Please note that the template includes reference to the NIH-supported data storage device that may or may not be suitable for your specific grant proposal. Contact Duke Research Computing (rescomputing@duke.edu) for more information on the use of that specific storage system. Updated April 2017.

  • Virtual Machines (VMs) Grant Templates
    Virtual Machines (VMs) – Virtual Computing Lab (VCL) & VM-Manage (template formats: TXT, RTF, Microsoft Word DOC, PDF). VCL and VM-Manage virtual machines are procurable through a web interface. VCL machines may be used for brief periods — usually between one and eight hours — and the machine is completely reset and wiped after use. VM-Manage virtual machines, which currently offer the Linux operating system, can be reserved for a complete semester, with a renewal option. Unlike VCL machines, VM-Manage machines persist and their states can be saved and even transported to other environments, such as a larger virtual machine or a […]
    Read More >>
  • Duke Compute Cluster
    Duke Compute Cluster is a service and support center that connects Duke University researchers to hardware, software, educational and personnel resources, both local and global, to enable novel computational science. We will leverage the parallel computing facilities already in place at Duke, help build out the computational infrastructure to handle future work-loads, foster the development of scalable applications and assist in the training of parallel-aware researchers. The Duke Compute Cluster works with OIT to manage and maintain a high-performance Linux cluster with nearly 680 machines and 7200 CPU-cores. The Duke Compute Cluster is based on a sharing model where interested researchers may purchase machines […]
    Read More >>
  • For Grant Applications
    For detailed materials to use in grant applications, please go here: https://rc.duke.edu/grant-resources/
    Read More >>
  • Letters of support
    Duke Research Computing routinely provides letters of support for NSF and NIH grant proposals in order to show institutional support for research proposals. Letters of support outline services that Duke Research Computing will offer, and staff are eager to discuss proposals and their computing, data storage, and data management requirements well before deadlines approach. Staff will also assist with detailed information for budget justifications relating to information technology. Contact rescomputing@duke.edu for additional information about obtaining a letter of support, assistance with grant preparation and budget justifications, or to get more information on any of the items on this page.
    Read More >>