Goal IV. Provide specialized support, appropriate facilities, and enhanced infrastructure for advanced research and scholarship

To meet the unique technical and support needs of persons who conduct computation-intensive research, the University provides high-end hardware, software, storage, and support, as well as large-capacity bandwidth programs. ITC offers targeted information technology resources and programs to help researchers accomplish their work and to advance the University's mission as a locus of knowledge. The 2005-06 academic year was an especially productive one in furthering IT-intensive research. Staff in both ITC and the Library have been evaluating the University's human and technical resources devoted to supporting computational research. They spent much time during the year discussing their findings and a set of proposals with deans and others. Another ITC-Library collaboration is resulting in new research computing support magnet centers to facilitate the use of information technology in research and scholarship. The following section describes those programs and others. The new large-capacity National LambdaRail network supporting research activities is described in Goal I, and other bandwidth issues are detailed in Goal V.

Research Computing Support Group

The Research Computing Support Group provides a wide range of services to advance the work of faculty and students conducting computation-intensive research and scholarship. Staff members provide educational outreach programs and doctoral-level technical assistance, with expertise in such areas as statistical computing, visualization, mathematical computing, and data access and archiving. Members of the Research Computing Support Group are responsible for acquiring and managing the University's high-end computational platforms, which are described later in this section, and they assist faculty and graduate student researchers on the use of those platforms. Staff members also facilitate researchers' use of high-performance computers that are available through collaborations with national supercomputing centers.

The group's outreach efforts include presentations, workshops, and electronic publications. Their monthly online newsletter, "Research Computing News," covers general IT information of interest to researchers, as well as updates about ITC services, software, and security issues. Presentations range from short informational sessions to in-depth workshops, and many of them are described under Research Computing Colloquia, below. Research group staff also offer specially tailored workshops on any of ITC's supported software packages.

Research Computing Support Magnet Centers

In summer 2006, ITC's Research Computing Support Group is relocating to two libraries. The move is designed to better serve the needs of researchers by leveraging the expertise of both ITC and Library research support staff. These new support areas are intended as "magnets" to draw scholars of widely varied backgrounds and abilities into new collaborative relationships with each other and with technical experts in the Library and ITC. The current Research Computing Support Center space in Wilson Hall is being converted to an ITC training classroom.

Part of the research group staff is moving to the Brown Science and Engineering Library in Clark Hall, where they will offer advanced science and engineering research computing support. The move is a merger between ITC's research support staff and the library's Digital Lab, creating a new digital Research Computing Lab. Other ITC research support staff will move to Alderman Library, where they will work with the Library's Digital Research and Instructional Services Group. The new facility, called the Scholars Lab, combines ITC research support with the Library's GeoStat and EText centers to provide advanced science and humanities research and instructional support.

Research Computing Colloquia

Research Computing Support Center staff continued their long-term educational outreach by hosting a technical colloquia series. Presentations are open to the entire University community and are held both at the center and elsewhere on the Grounds. Presentations are posted on the center's Web site for reference. Among the 15 presentations offered during 2005-06 were a hands-on high-performance cluster tutorial and talks on effective display of quantitative data, parallel computing, and the use of such applications as LaTeX, SAS, and Matlab. In addition, research center staff offered in early fall a comprehensive overview of research hardware, software, data storage, and support services available at the University.

Framework for Supporting Digital Scholarship and Research

Staff from ITC and the Library have been collaborating since fall 2005 to create a proposal for an institution-wide framework for supporting digital scholarship. The draft report, entitled "A Technology Framework for Supporting Digital Scholarship and Research," evaluates current resources currently available across the Grounds to support the development and conservation of digital scholarship. It argues for the adoption of an institutional framework to support the full range of activities encountered by scholars in developing their instructional and scholarly digital projects. The report also includes recommended enhancements to the University's technical infrastructure, technology and support services, and pertinent policies. During spring 2006, ITC and Library staff made presentations about their proposals to various groups and academic departments.

High-performance Computing Platforms

During 2005-06, ITC continued to acquire the high-performance computing platforms that underpin a great deal of research and scholarship at the University. The newest cluster is the largest to date at U.Va. and replaces the first 48-node Linux cluster installed fall 2002. The new cluster nearly doubles the total number of high-performance research computing nodes from approximately 220 to nearly 375. While this represents a vast improvement in computing resources for researchers, the University continues to lag behind peer institutions in hardware, support staff, and facilities dedicated to research computing and computational science.

The newest cluster joins a 125-node cluster acquired in 2005 and a 38-node cluster that has been used since 2003. The oldest Linux clusters are heavily used for parallel and serial computing, and they have been updated with new operating systems and additional security. In addition to the Linux clusters, ITC provides Sun clusters, Sun workstations, an SGI cluster, and an IBM p640 Model B80 Symmetric Multi-Processor (SMP), which serves as a bridge system for researchers whose local computational needs in software or libraries are tied to the AIX operating system. The IBM SMP serves as a separate high-performance computing platform and has been configured to use the same batch queuing system as ITC's Linux clusters.

In spring 2006, University researchers were offered the chance to purchase nodes in the new Linux cluster for their use. Among the benefits are priority access to the purchased nodes, free system administration, timely security updates, constant monitoring, and the ability to use more nodes than purchased during periods of non-peak demand.

Research Computing Infrastructure Task Force

During spring 2006, a pan-University faculty-led task force drafted a report that describes the pressing needs for hardware, staff, and space to advance computation-intensive research in engineering and the natural sciences, as well as in the social sciences, education, and humanities. The task force was formed in response to calls by University leaders for the institution to achieve world-class status in research and education in the computational sciences. The report not only details areas in which the University lags behind its peers in current technologies and support, but also it includes specific cases in which investments in new technologies will dramatically improve research and scholarship. In developing the report, the task force sought specific descriptions from faculty about how they will use world-class computational technologies and support to advance their scholarship.

Staff from ITC's Research Computing Support Group participated on the task force, as well as on development of a proposal to the National Science Foundation that emerged from the task force's work. The proposal is for a computational science instructional initiative, which will include ITC staff members if it is funded.

Hierarchical Storage Management

For the past seven years, ITC's Hierarchical Storage Management system (HSM) has provided storage to faculty and staff for large and infrequently used data. Although the system is not intended for files that are accessed often, users may name, copy, and retrieve needed files from their own workstations. Files on the HSM, which is managed by ITC's Unix Systems Group, are stored on magnetic tape and backed up nightly. Use of the HSM grows annually. During 2005-06, the amount of user data stored in the system grew by 14 percent, while the number of users remained constant.

During the past year, ITC architected and planned a trickle-down approach to help solve a combination of resource problems. The key aspect of this combined effort is the migration of the Research Data Hierarchical Storage Manager to the new storage infrastructure, described below. Not only will this improve performance for researchers, but also it will free up older tape equipment for use on U.Va.'s mainframe. This, in turn, will enable a large amount of mainframe tape consolidation and free the computer room floor space needed to hold the next research computing cluster.

New Storage Services

ITC, working with Alderman Library, made available a new large-scale hierarchical storage management system that will be used to store the library's critical digital assets. The first departments taking advantage of the system are Alderman Library's Digital Library Production Services and Rare Materials Digital Services. Those departments are archiving thousands of single-master copies of DVDs containing digital material. The technology selected for this deployment is designed to integrate into and support the joint ITC/Library vision for an enterprise-wide storage infrastructure that can facilitate digital scholarship and still meet the needs of other applications and services across the enterprise.

Academic Computing for Health Sciences

For the past 13 years, Information Technology and Communications-Academic Computing Health Sciences (ITC-ACHS) has provided high-end computing support to biomedical researchers in the Health System and across the Grounds. Services include support for image acquisition and processing, bioinformatics, high-performance (cluster) computing, and bioinformatics database programming. The client base includes approximately 450 faculty, staff, and students in more than 100 departments and centers. The facility is available 24 hours a day.

Biomedical researchers at U.Va. are using target databases that continue to increase exponentially in size. The number of computing jobs on ACHS's cluster and the average length of each job are increasing as well. This year, ACHS expanded its computing cluster to 108 processors, 180 gigabytes of RAM, and 8 terabytes of disk space. Approximately half a million jobs were run by and for users. Because existing cluster queuing systems will fail if more than a thousand jobs are queued, ACHS staff developed a second-tier queuing system that selects the highest priority jobs from a database and sends them incrementally to the conventional queuing system. More than 250 users access locally maintained copies of GenBank and UniPROT. Disk capacity on the facility's GCG server was recently increased to 4 terabytes.

ACHS staff assisted during the past year with two new projects of particular interest. The first was a drug-screening project by Drs. Jae Lee and Dan Theodorescu, in which the responses by cancer cells to approximately 50,000 putative drugs were analyzed. The second, working with Dr. Mike Timko in the Biology Department, was validation of 150,000 clones from the cowpea genome. ACHS now houses the world's only publicly accessible cowpea gene database. The cowpea, or black-eyed pea, is a critical food source for 200 million people in west and south Africa. Work continues on implementation of protein identification algorithms for the Cell Migration Consortium and on parallel Mass Spectroscopy algorithms. ACHS also continues to administer the "GEOSS" gene expression LIMS and archiving system (developed at ACHS). Staff are starting work on a custom LIMS system for the Keck Biomedical Mass Spectrometry Laboratory.

ACHS's Data Acquisition and Multimedia labs continue to be well utilized. Lab staff logged more than 1,500 client contacts during 2005-06. The lab's new podcasting provides just-in-time delivery of hints on common Photoshop questions, and has been very well received by users. Because lab users also are bringing and generating ever-larger data sets, the "Meduser" data store was increased to 1.2 terabytes. Meduser allows storage of users' files for up to three weeks. Finally, increasing demand for video editing capabilities necessitated substantial upgrades during the year to ACHS's video capture and rendering systems.

Cancer Center

Since 2002, staff in ITC's Departmental Applications Development (DAD) Group have refined, enlarged, and maintained the database that U.Va.'s Cancer Center staff use to track information related to medications, treatment, and quality of life. The Investigational Pharmacy, which is affiliated with the Cancer Center, tracks results of studies of new medications for cancer treatment. The database now includes a Web interface to enroll patients from other U.Va. medical departments and other medical facilities nationwide. ITC staff continue to work on the programming and infrastructure needed to enable multi-center trials for the Cancer Center. Enhancements and specially requested reports are completed as needed.

In a closely related area, Craig Slingluff, M.D., of the Human Immune Therapy Center has requested ITC's assistance in combining his older research databases into a single relational database to assist his research team. DAD staff are collaborating with the Public Health Sciences Department on this effort.

National Partnership for Advanced Computing Infrastructure

For the past several years, University researchers have had access to dozens of high-end computational platforms and software at remote locations through the institution's participation in the National Partnership for Advanced Computing Infrastructure (NPACI). A partnership of more than 40 leading research universities and several corporations, NPACI started with funding from the National Science Foundation in 1997 to create a national, distributed computing environment that is seamless and secure. U.Va. is a research and education partner in the program, and faculty researchers offer expertise in the development of metasystems, programming tools, and environments. ITC staff coordinate the use of NPACI resources by faculty and staff.

Goal IV Appendix