A comprehensive plan for information technology
in support of research and scholarly accomplishment at IU
On this page:
- RAC Organization and Services
- Collaboration, Communication, and Interactive Computing
- Massive Data Storage Initiative
- Digital Libraries
Christopher S. Peebles
Associate Vice President, Research and Academic Computing
Dean, Information Technology (Bloomington)
Craig A. Stewart
Director, Research and Academic Computing
Director, Research and Academic Computing
Donald F. McMullen
Principal Scientist, Advanced Information Technology Laboratory
Senior UITS Staff1
So, computers cannot create information.
This has long been clear.
What has only recently become clear is that there
is no such barrier to computers creating knowledge
Indiana University President Myles Brand has set an exciting and challenging goal for the University: to become a leader, in absolute terms, in the use of Information Technology.This mandate was given to Michael McRobbie when he assumed the duties of Vice President for Information Technology and CIO in January, 1997. He in turn made the quest for IT leadership a cornerstone of the 1998 reorganization of IU's information technology organizations into University Information Technology Services (UITS). The Indiana University Information Technology Strategic Plan: Architecture for the 21st Century3 followed shortly thereafter and described a strategy for IU to become a leader in the use of information technology in its scholarly, strategic, and business activities. This Strategic Plan was developed by representatives drawn from the entire Indiana University community - faculty, staff, and students from all eight IU campuses -and has been presented to and endorsed by student and faculty advisory groups and councils on all of these campuses.
The IU IT Strategic Plan places great emphasis on the role of information technology in research and the academic pursuits of the University. Innovative and useful research is, after all, a key component in all of the missions of Indiana University. Information technology can add significant value to pure and applied research and to the development of students and their accomplishments in the sciences, the arts, and the humanities.
The use of computation in research has a long (60-year) and productive history in government, in industry, and especially in universities and other institutions of higher learning and research. The automation of calculations by digital, electronic computers,which all but eliminates random errors due to human frailties and fatigue,has played a major part in the solution of recondite equations in nuclear physics, calculated trajectories for artillery shells and orbits for the earliest space flights, enabled the breaking of German ciphers in World War II, and led to the success of the Human Genome Project. The speed of these calculations, which compress human-scale efforts into nanoseconds, permits the magnitude of mathematical processes to transcend what might be accomplished in the coordinated but unaided efforts of millions of human lifetimes. For example, the recent (1993) calculation of the rest mass of the proton by Don Weingarten (formerly professor of physics at IU) and a team from IBM required a computer constructed especially for the task and took approximately " ...one hundred quadrillion(1017) arithmetic operations,"4 a number that taxes even the most expansive imagination.
The use of computation and the other elements of IT have spread far beyond complex arithmetic operations and the physical sciences. The storage, transformation, evaluation, and selection of massive amounts of digital data are crucial to research in biology, physics, comparative literature, business, art history, and anthropology. The Human Genome Project, the Advanced Photon Light Source, and the daily output from point-of-sale devices in retail stores can produce thousands of billions of bytes (PB, petabytes) of data in a relatively short span of time. These data must be safely stored, sorted, transmitted, transformed,and presented so that they can serve to illuminate the structure of the genome, the behavior of condensed matter, and the choices made by consumers. They can thereby help with the understanding of the genetics of health and disease,5 fundamental physical processes, and enterprise resource planning for large retail chain stores.6 The increasing size of computational problems has led to the linkage of geographically dispersed high performance computers and storage devices through high speed digital networks into national and international computational grids. Likewise, the dispersion of instruments, scholars, and primary sources of information have led to virtual laboratories, collaboratories, and communities of practice made possible by these same networks.7 In many ways research on the frontiers of biological and physical sciences is tied to progress in what has been called the "third science"8 - high performance computing, with its ability to simulate complex natural processes in the fine and in the large. Much work remains to be done beyond these frontiers in physics and biology; there is much promise in the application of high performance computing to the work of scholars in the arts, humanities, health services, social sciences, and engineering.Thus the provision of academic computing and networking services must encompass not only high performance computing (HPC) but high bandwidth networking, high volume storage systems, and high impact organizations that support the scholarly use of these IT facilities and services. We have become, as Fritz Machlup pointed out nearly two decades ago, a society of informavores9 who live in an intellectual ecology near the edge of chaos. The Indiana University Information Technology Strategic Plan directly addresses the needs and requirements of education and research in this rapidly changing world and speaks to the challenges that face the entire University, including UITS.This document sets forth detailed plans for UITS to achieve the broad and far-reaching goals set for research and academic computing within the IU Information Technology Strategic Plan.
The delivery of IT services in support of science and scholarship at Indiana University falls largely to the Research and Academic Computing Division (RAC) of University Information Technology Services. Of the ten key recommendations of the Indiana University Information Technology Strategic Plan, two directly address responsibilities of the RAC organization.
Recommendation 5. In support of research, UITS should provide broad support for basic collaboration technologies and begin implementing more advanced technologies. UITS should provide advanced data storage and management services to researchers. The University should continue its commitment to high performance computing and computation, so as to contribute to and benefit from initiatives to develop a national computational grid.
Recommendation 9. The University should build upon and expand its digital library program,and develop the digital library infrastructure needed to support research,teaching and learning.
These two major Recommendations, and the Actions which flow from them, provide strong direction for future investments and ongoing efforts of those who make up the RAC organization. These Actions, and the organization of RAC itself for that matter, can be grouped in four broad categories: advanced computation, collaboration,massive data storage, and digital libraries. Indiana University, UITS, and RAC have strong intellectual and material foundations in each of these areas, resulting in notable accomplishments over the last several years.The leadership, intellectual capital, and track record of UITS and its faculty partners positions IU to achieve and sustain a national leadership role in research computing for the foreseeable future.
This document begins with an analysis of recent historical trends in the provision of services in the research and academic computing areas. It then defines the present organizational structure and activities of the Research and Academic Computing Division of UITS. Next, it enumerates the 17 Actions that follow from the two broad Recommendations quoted above and describes the implementation strategy and immediate implementation tasks for each Action (or small group of related Actions). In this way, this document explains the steps that the RAC Division of UITS is taking, and will continue to take, over the remaining years of the five-year strategic plan process, to assist Indiana University in achieving the goal set out by President Brand to become a leader, in absolute terms, in use of Information Technology. With such an ambitious goal it is essential to carefully measure and chart progress toward goals.This document ends with a summary of involvement by RAC in the UITS Quality Initiative and a brief assessment of the current state of RAC services.
Indiana University has established itself nationally as a leader in high performance computing (HPC). Through successful competition for government and private funds, IU scholars and UITS have acquired two supercomputers that are listed among the 500 most powerful computers in the world.10 UITS system engineers have demonstrated their intellectual leadership in HPC through the development of a PC cluster for parallel computing - one of the first such clusters in the US able to run both of the major operating systems for Intel hardware. This leadership extends beyond the acquisition and operation of HPC systems, however. IU computer scientists are nationally and internationally recognized as leaders who will shape the future of high performance computing. Using the HPC facilities provided by UITS, IU researchers have twice won HPC Challenge awards at the annual IEEE/ACM Supercomputing conferences.11,12 The use of HPC systems at IU has enabled innovative research in such disciplines as astronomy, chemistry, physics, biology, and mathematics. Furthermore, the use of high performance computers combined with virtual reality facilities in the Advanced Visualization Laboratory(AVL) have enabled researchers in several departments to create 3Drepresentations of their research data and of their artistic creations. The Department of Theatre and Drama, for example, partnered with AVL to produce some of the largest photorealistic simulations of complex lighting systems ever produced.
- measured in Terabytes
Indiana University has made significant investments in HPC hardware in recent years, and the results of these investments are shown graphically in Figure 1. Increased investment will be required to maintain a leadership position. This investment must come in the form of increased support staff as well as purchase of newer and faster systems.
HPC and advanced visualization technologies are now heavily used by dozens of IU researchers on the Bloomington and Indianapolis campuses. That number will increase by hundreds of scholars encompassing all IU campuses in the not-too-distant future. Through application of the principles of the IU Leveraged Support Model,13 RAC plans to increase use of HPC technology across all eight IU campuses. Through innovative research in computer science, applications of computational science, and research partnerships with industry, IU can continue to lead in the deployment and use of the world's most powerful computers.
The use of research computing, however,is not limited to the hundreds of scholars who use HPC resources. Rather,it is built into the syllabus of courses in every school. It plays an important part in courses on mathematical modeling in the Kelley School of Business; it is integral to the curricula in computer science, chemistry,physics, and engineering. In 1999, the UITS User Survey14 showed that 47.5% of the Bloomington faculty, undergraduate students, and graduate students used RAC services, and on the Indianapolis campus 13.8% of the faculty and graduate students used these services. This document describes ways in which UITS can enhance services provided to current users and expand use of RAC services to include a larger portion of the University community.
Collaboration, Communication, and Interactive Computing
The research community at Indiana University includes a distinguished group of scholars who develop and work with collaborative tools. For example, scholars in several IU schools and departments have participated in some of the largest experiments of collaborative virtual environments ever undertaken, including a multinational demonstration of NICE15 (Narrative, Immersive, Constructionist/Collaborative Environments) at SuperComputing '97. Collaboration via the use of advanced IT facilities has become crucial to the strategic use of IT for research, scholarship,and learning. UITS must simultaneously pursue the implementation of collaborative technologies that are available for use today as well as conduct research in the most advanced areas of immersive, collaborative VR(virtual reality) technology that will sustain scholarship and learning into the future.
The sophisticated use of visualization,virtual reality, and collaboration technologies by the IU research community can also be used to enhance work in other areas, such as e-commerce and distributed education. Researchers in the IU Kelley School of Business are engaged in groundbreaking efforts to develop e-commerce applications based on virtual shopping environments as well as collaborative environments for teaching and learning about e-commerce. In addition, the University has an opportunity to improve the learning process across disciplines, enable greater opportunities for distributed education,and reduce costs and risks through use of simulation-based laboratory exercises.
Massive Data Storage
Accomplishments resulting from IU's Massive Data Storage Initiative have won national and international attention for the University and for UITS. This initiative permits physicists at IU to store and analyze a 5 TB data set from one of the most sophisticated high-energy physics experiments conducted to date. At the same time, this initiative will be implemented in a manner that ensures these technologies offer value to everyone at IU. HPSS (High Performance Storage System) is one of the core technologies of the Massive Data Storage Initiative. HPSS is designed, developed, and produced by a consortium that includes the largest university and research institutions in the world. Intellectual leadership by UITS in the field of massive data management is demonstrated by IU's membership in the HPSS consortium and the lead role that UITS has achieved in developing value-added tools for HPSS. Some of these tools will enhance the ability of IU researchers to search through and analyze extremely large data sets. Other tools under development at IU include a highly intuitive user interface for the massive data storage system. This interface is a key component in efforts by UITS to assure widespread use of the massive data storage system.
HPSS and software developed at IU have made it possible to manage extremely large quantities of data. Investment in tape robots and media has greatly increased the amount of data storage available within IU facilities. The growth in capacity of UITS tape storage system is shown in Figure 2. Thanks to the Massive Data Storage Initiative,and the concomitant increase in ability to manage data, researchers will be able to analyze data from experiments and perform computations in ways that would otherwise be impossible.
- measured in Terabytes
Figure 3 displays the recent growth in UITS disk storage systems. This growth has occurred in concert with growth in tape storage, and is part and parcel of UITS facility enhancements for dealing with massive quantities of data. However, the current set of disk storage systems is spread across multiple computing systems using different - and often incompatible - storage protocols. A transparent and robust common file system is an essential part of the infrastructure necessary for effective electronic collaboration by the IU community. This document sets forth a blueprint for a dramatically improved set of UITS central file services, building upon the continued development of the infrastructure of the Massive Data Storage Initiative.
- measured in Terabytes
Indiana University has long been a leader in several areas of digital library activities. The VARIATIONS16 project, which provides a multimedia interface to music and delivers CD-quality sound, is nationally recognized as "best of breed" in multimedia music library systems. It is one of the tools used by the Indiana University School of Music to maintain its position as one of the best educators of current and future leading musical performers, educators, and theorists. Other exemplary projects include the Hoagy Carmichael Collection,17 the Hohenberger Photograph Collection18 of southern Indiana subjects and scenes, and the digital photography collection of the Gary Works, US Steel.19 The Library Electronic Text Resource Service (LETRS)20 and closely related Victorian Women Writers Project21 are the unquestioned leaders in production and delivery of electronic texts in the humanities. In fact, in April 2000, the Victorian Women Writers Web site was accessed more than 40,000 times by scholars from throughout the world.
These and other initiatives are carried out with the support of UITS but within the structure of the IU Digital Library Program, which is a collaborative effort of the Indiana University Libraries, the Office of the Vice President for Information Technology,UITS, and the School of Library and Information Science. This collaboration capitalizes on the institutional capabilities of the University. It focuses University resources on digital library projects that support the teaching and research of IU faculty, supports the learning and research of IU students, and fosters digital library research in general. Digital libraries will be increasingly critical to the University's intellectual and operational activities. This document describes a comprehensive plan for implementing a broad spectrum of digital library activities that will be critical to the University's achievement as we enter the new century.
Leadership by Indiana University in many areas of information technology, and its commitment to excellence in IT across the board, is fundamental to sustaining leadership in research computing and digital libraries. Both areas are bolstered by the international profile in high performance networking enjoyed by Indiana University. The Massive Data Storage Initiative is made possible by a combination of elements: the vision of the University's leadership, accomplishments in University Information Systems, and achievements in research computing. The interlinked efforts of Research and Academic Computing and the Teaching and Learning Information Technologies Division, itself a leader in the use of IT for educational purposes, underscore the interdependence of teaching, learning, and research in philosophy and practice. The success of the parts feeds into the success of the whole.
RAC Organization and Services
The Research and Academic Computing Division of UITS is organized and funded to produce a comprehensive suite of IT services and support for scholars and students on the Bloomington(IUB) and Indianapolis (IUPUI) campuses of Indiana University. RAC personnel also offer support for scholars and students on the other campuses of Indiana University who use the central computing and storage services offered by UITS. The three RAC subunits, outlined below, report to Associate Vice President for Research and Academic Computing Christopher S. Peebles, who is also Dean for Information Technology on the Bloomington Campus.
Detailed information about all RAC services and activities is available via the Web at http://uits.iu.edu/ . If you have questions or comments, you are invited to send e-mail to rac (at) iu.edu.
The computation component of RAC, directed by Craig Stewart, has one production and four education and support units.
Research and Technical Services, led by Mary Papakhian, deploys, maintains, and develops the HPC environment for IU. The personnel in this unit are responsible for the IBM RS/6000 SP, Sun E10000, the Compaq PC cluster, and central Unix servers for research and instructional use. They are based on the Bloomington campus, although the services (cycles and applications) are available to scholars on all campuses via IUNet.
The Center for Statistical and Mathematical Computing (Stat/Math Center), managed by John Samuel, supports the use of mathematical and statistical applications by students, staff, and faculty throughout the University. The Stat/Math Center is a component of the Interdisciplinary Consortium for Statistical Applications, a partnership among UITS, Research and the University Graduate School (RUGS), and the Department of Mathematics on the Bloomington Campus. The Center has presence on the Bloomington and Indianapolis campuses, and, as a RUGS Center, offers its services to all campuses of the University.
The High Performance Computing Support Team (HPCST), managed by David Hart, is based in Bloomington and is responsible for University-wide support of parallel computing and program libraries. The HPCST plays a critical role at IU in facilitating the use of IU's sophisticated HPC systems.
The Computational Research Support Group (CRSG), managed by John Daniels, is based in Indianapolis and is responsible for University-wide support of data mining and bioinformatics software; this team also provides local support for parallel computing to the IUPUI campus. The personnel in HPCST and CRSG work directly with students and faculty who use the IBM RS/6000 SP and the Sun E10000 supercomputers and the Compaq PC Cluster facilities in their research.
The Unix Systems Support Group(USSG) provides support for individuals, departments, and schools that use Unix servers and workstations as part of their academic pursuits. This group has personnel on the IUB and IUPUI campuses.
RAC-Data and Presentation
The data and presentation component of RAC, directed by Gerry Bernbom, has three chief areas of responsibility.
The Distributed Storage Services Group (DSSG), managed by Anurag Shankar, is responsible for extensive disk and automated tape facilities on the Indianapolis and Bloomington campuses. Their efforts and intellect, along with those of Gustav Meglicki from the Office of the Vice President for Information Technologies, comprise the IU contribution to the HPSS Consortium.
The Advanced Visualization Laboratory (AVL), managed by Eric Wernert, provides a CAVE 3D virtual reality space on the Bloomington Campus and an Immersdesk facility on the Indianapolis campus. The engineers who maintain these facilities also support faculty and students in their use. Projects that achieve visual representation in these virtual and immersive environments range from astrophysics to medical images to molecular structures.
The Library Electronic Text Resource Service (LETRS) is a partnership between UITS and Indiana University Libraries. Managed by John Walsh, its personnel support the acquisition, creation, and access to digital text collections by faculty,students, and staff. LETRS also provides part of the UITS contribution to the Indiana University Digital Library Program.
Advanced Information Technology Laboratory(AITL) - High Performance Network Applications.
Rick McMullen directs the Advanced Information Technology Laboratory and is also Principal Scientist with responsibility for High Performance Network Applications. The AITL offers an environment for the testing and development of cutting-edge network applications. At the moment, AITL engineers assist and coordinate projects supported by the IU High Performance Network Applications Initiative. They also assist in the development and testing for Internet2 applications for the Abilene Network.
High Performance Computing and National Technology Grids (Actions 29 and 31)
ACTION 29. In order to maintain its position of leadership in the constantly changing field of high-performance computing, the University should plan to continuously upgrade and replace its high-performance computing facilities to keep them at a level that satisfies the increasing demand for computational power.
ACTION 31. The University should plan to evolve its high-performance computing and communications infrastructure so it has the features to be compatible with and can participate in the emerging national computational grid.
UITS HPC resources include a139-processor IBM RS/6000 SPTM, a 64-processor Sun E10000TM, and a 64-processor cluster of Compaq ProLiantsTM. IU and UITS enjoy a strong partnership with IBM, based on exchange of research ideas and hardware granted to IU through IBM's Shared University Research (SUR) grant program. IU and UITS also enjoy a strong partnership with Sun Microsystems. Sun recently named IU a "Center of Excellence" in high performance computing and partners with IU in research projects. Similarly, Compaq and IU have partnered in the development of IU's parallel PC cluster. All of these relationships enhance the range and quality of services offered by RAC.
Because of the dynamic nature of the traditional supercomputer market, it is essential for UITS to maintain two large traditional supercomputers from mainstream vendors, as well as a parallel computing cluster. Given current architectural trends, maintaining one distributed-memory system (currently the IBM RS/6000 SP), one shared-memory system (currently the Sun E10000), and a cluster of commodity PCs is the best choice for the immediate future. IU must, however, constantly reanalyze the market to ensure that the University community has access to the best possible resources. To enable the University to remain a leader in the delivery of HPC systems, UITS must leverage its own resources with external funding and corporate partnerships. By doing so, UITS can assure that IU researchers are the first to perform massive new calculations that push back the frontiers of human knowledge with ever-increasing speed.
Two other emerging trends that require action by UITS and the University include grid-based computing and the use of diffuse resources for high-throughput computing. Grid-based computing is one of the most important trends in HPC today. High performance computers are tied together by high-speed networks to create resources for solving problems that are larger than any individual computer could address. Massive storage systems, high-output instruments, and visualization facilities are tied into such grids to create problem-solving and scientific discovery facilities of heretofore unheard of capabilities.Grid-based computing is becoming a reality, and Indiana University has already established a significant place in this field. UITS must lead efforts to make grid-based computing an everyday asset for researchers throughout IU. The second emerging trend is the use of diffuse resources,such as employing underutilized desktop computers, to perform high-throughput computing.22 This approach is important when scientific discovery requires the solution of a very large number of basically similar problems. This need is quite common in the natural and physical sciences. The technology for harnessing unused cycles available throughout the University IT infrastructure, long just a computer science research project, is nearing readiness for pilot use, and should be ready for routine production use in one to three years. By adopting these technologies, IU can leverage its existing hardware infrastructure in a cost-effective fashion.
- Through IBM Shared University Research grants and IU investment,
continue to expand and upgrade the research-use IBM RS/6000
- Ensure that Indiana University has local access to the best
available shared-memory supercomputer, which at the moment is
- Through partnerships, grants, and IU investment, continue
to expand the presence of HPC resources built around clusters
of commodity components. Investigate the possibility of expanding
existing resources to include a cluster based on the Compaq
- Test, trial, pilot, and implement software systems that permit
use of installed microcomputers and workstations throughout
the University to perform high-throughput computing.
- Continue and expand efforts to make grid-based computing an
everyday asset for researchers at IU. In particular, lead efforts
to create a CIC-wide computational grid, and continue to partner
with NCSA (National Computational Science Alliance) and NPACI
(National Partnership for Advanced Computational Infrastructure).
- Continue participation in the Globus computational testbed
and other national grid development projects.
- Enhance local support for high performance computing, grid-based computing, and high-throughput computing.
Computation- and Information-Intensive Applications and Research Software (Actions 30 and 33)
ACTION 30. The University needs to provide facilities and support for computationally and data-intensive research, for non-traditional areas such as the arts and humanities, as well as for the more traditional areas of scientific computation.
ACTION 33. The University, through UITS, should provide support for a wider range of research software including database systems, text-based and text-markup tools, scientific text processing systems, and software for statistical analysis. UITS should investigate the possibilities for enterprise-wide agreements for software acquisitions similar to the Microsoft Enterprise License Agreement.
Actions 30 and 33 speak to a diverse set of needs and diverse groups within the IU community. These groups have in common the application of quantitative changes in computing power to create qualitative changes in the their artistic and scientific accomplishments. Many of the techniques developed in the high performance computing and communications community are expanding in usefulness to an increasingly large segment of the academic and artistic community at IU. The capabilities of software applicable to the social and natural sciences has increased dramatically, for example in the areas of data mining and bioinformatics. At the same time, the University's academic strengths in the arts, humanities, and health science stand to benefit from the application of better software and faster hardware in such fields as visualization and simulation. With improved access to hardware, software,and support, IU can significantly increase its scientific, clinical, and artistic productivity in a wide variety of disciplines. UITS strategic objectives related to Actions 30 and 33 can be subdivided into three areas: support for computationally-intensive research, central facilities for computationally-intensive applications, and access to research software.
Enhance Support for Computationally-intensive Research
Within some areas of interest and on some campuses, support for computationally-intensive research has been delivered in a fashion that is both well regarded on campus and recognized nationally for its excellence. The services provided by the Stat/Math Center and Unix Workstation Support Group on the Bloomington campus are two examples. IUB researchers who use cutting-edge parallel programming techniques have also received good support. Broadly considered, however,support for computationally-intensive research has been delivered inconsistently across campuses and topic areas. Professional consulting assistance for most areas of computationally-intensive research applications on the IUPUI campus has historically been inadequate. Consulting assistance and ease of access to IU's largest research systems for researchers on IU regional campuses has likewise been poor.Some types of research computing have been inadequately staffed generally.For example, important emerging software areas, such as bioinformatics, are not yet supported.
With the best HPC facilities of any university in the state, IU must ensure that researchers at all campuses receive the benefits of these excellent resources. Thus, staffing on the IUPUI campus must be enhanced, particularly in support of research application software. This should enable development of deep and ongoing customer/client relationships that will enable IUPUI-based researchers to make the best use of the University's computing resources. Likewise, RAC staff who support scholars on the regional campuses should be constantly available to researchers by phone and e-mail, and for meetings in person at the regional campuses.
The tools in support of high performance computing, grid-based computing, and high-throughput computing are advancing rapidly. Application of the IU leveraged support model to this area will permit provision of online help 24 hours a day, 7 days a week. In this way, IU can reap the greatest benefit from the expertise of its technical support staff and systems engineers, and provide the tools and information that will help computer users solve many of their own problems.
- Enhance support, and thus accessibility, of high performance
computing resources. Specifically, RAC will enhance use of these
resources by communities who could benefit from these systems,
but are not now using them extensively. RAC will target researchers
on the IUPUI campus and the six regional campuses, as well as
researchers at IUB who have not yet adopted high performance
computing techniques in their research.
- Enhance support for computation-intensive and high performance
computing, including data mining and bioinformatics applications.
- Enhance support for, and thus utility of, high-powered Unix workstations at the regional campuses and at IUPUI.
- Create a research support group, located in Indianapolis,
to provide easily accessible support for researchers at IUPUI
and the six regional campuses. Through this group, provide leadership
in support of emerging computationally-intensive computer applications.
(This task is partially complete with the formation of the Computational
Research Support Group,which consists of a manager with expertise
in data mining, an HPC specialist, and a bioinformatics/statistics
support specialist. The first two of these positions have been
filled, and a search is currently underway for the bioinformatics
- Expand the IUB-based High Performance Computing Support Team
(HPCST),in part to provide more resources for support of IU's
regional campuses.(This group was expanded to include an additional
user support specialist position 3Q 1999.)
- Schedule time for members of the Stat/Math Center to be available
for consultation on the IUPUI campus. (Implemented 2Q 1998.)
- Enhance support for Unix workstation use. Add a user support specialist (Unix workstation support) to be physically located at IUPUI.This staff person will serve primarily as a support resource for users of Unix workstations on the six regional campuses, but will also provide some support for users of this technology at IUPUI. (Completed 3Q2000.)
Enhance Central Facilities for Computationally-Intensive Applications
The criticism that high performance computing was limited in its utility to high-energy physicists, astronomers, and a smattering of chemists was once easy to level at supercomputing centers because it contained a large element of truth. However, computationally-intensive applications are now used in a broad array of sciences as well as the fine arts. Although many HPC applications in the arts and sciences can be used interactively, traditional supercomputers are most effective when used in a batch-processing manner. HPC systems at IU are used, for the most part, in traditional batch mode. For HPC to be of value for all areas of academic inquiry, UITS must also provide excellent facilities for highly interactive, easily visualized data- and computation-intensive applications. UITS must create computing resources that fit easily with the way social scientists, management scientists, medical scientists, environmental scientists, humanists, and artists practice their craft. That effort will require the addition of hardware and software designed for interactive,data-intensive computing.
- Create hardware and software environments specifically designed
to facilitate use of interactive computing applications in research.
Change system architectures and management facilities, and add
application software that will make these environments appealing
and useful to a larger group of researchers. Collaborate with
other UITS and RAC subunits, particularly AVL, to enhance facilities
for interactive visualization of computationally-intensive research
- Enhance accessibility of interactive computing on large central servers to undergraduates at IUB and especially IUPUI.
- Create a hardware and software environment specifically designed
for high-end interactive computing. (The recent upgrades to
the IBM RS/6000 SP and the acquisition of the Sun E10000 are
recent significant steps in this process.)
- Obtain and support software for new types of data-intensive
research, including data mining and bioinformatics. (SAS/Enterprise
Miner is slated for installation at IU during the fall 2000
- Promote the cluster of Sun servers (known as Steel) as the
primary resource for educational activities that make use of
research software,including database software. Over time, plan
for phase-out of the smaller and older IUPUI-based Auburn system.
Improve processes for gaining access to and using Steel at IUPUI
and at IU's regional campuses.
- Purchase software and hardware designed to facilitate use of the IBM RS/6000 SP, the Sun E10000, the Compaq-based parallel PC cluster, and immersive visualization environments for computational steering applications.
Enhance Access to Research Software
Access to software is a fundamental problem in research computing. This problem can be separated into two tiers:
- Everyday access to research software tools commonly used by
researchers. Statistical software is an obvious example. UITS
has long been successful in providing research software on campus
at prices well below academic list. Prior to the year 2000,
however,UITS had not been able to negotiate a license that provided
widespread access for home use of such tools by students. A
less obvious but still important issue is access to personal
productivity tools for users of Unix workstations. Many researchers
on campus do most of their work on a Unix workstation, but keep
a traditional microcomputer nearby for word processing and spreadsheet
- Affordable access to more specialized, but still widely used, research tools. Access to common statistical analysis tools is a clear need throughout the University. This need is tightly related to IU's goal of increased attention to quantitative thinking as part of IU's curricula, a goal explicitly established in the Indiana University Strategic Directions Charter.23 Affordable and easy access to research software enhances research and creative accomplishment by IU researchers, and is an area of significant achievement by UITS. This effort must be further expanded to include more types of research software. Throughout its history, UITS has fostered significant intellectual achievements and economies of scale via innovative site licenses for research software. UITS should continue to enhance its services and capitalize upon further opportunities to expand access to important research tools.
- Negotiate aggressive and highly beneficial site licenses for
the most commonly used research software tools, aiming for very
low cost barriers and home use privileges for students, staff,
- Continue to negotiate favorable terms for site licenses for
software resale on campus, aiming particularly to provide access
to a greater number of software tools for IU's regional campuses.
- Enhance accessibility of important niche software packages from UITS central servers. For many types of research software, the best and most sophisticated implementations are those created for the Windows NT operating system. Thin client technology permits the delivery of NT-based research software on non-NT systems, or on NT-based systems that lack copies of software licensed specifically to that system. Such flexibility is of great value because it can facilitate collaborative use of research software that is highly specialized, and in many cases extremely expensive.
- Obtain site licenses for statistical and mathematical software,
similar to the Microsoft Enterprise License Agreement. (Done
3Q 1999.) (UITS announced in February 2000 a site licenses for
use of SPSSTM throughout the University and on the
personally-owned computers of IU community members. As of 1
July 2000 over 1,400 copies of the software have been distributed).
- Obtain site licenses for Unix-based productivity suites.(ApplixwareTM
and Island SoftwareTM licenses signed and announced
- Create an NT-based server to deliver important but highly
specialized research software tools to the desktops of IU's
- Expand software available through UITS facilities, especially
data mining software and software of interest in biochemistry,
chemistry, and medicine.
- Continue to achieve cost savings and economies of scale through software resale efforts.
Collaboration (Actions 27 and 28)
ACTION 27. UITS should launch an aggressive program to systematically evaluate and deploy across the University state-of-the-art tools and infrastructure that can support collaboration within the University, nationally and globally.
ACTION 28. UITS should explore and deploy advanced and experimental collaborative technologies within the University's production information technology environment, first as prototypes and then, if successful, more broadly.
All researchers work in intellectual communities and communities of practice. A fundamental feature of this work is collaboration - across disciplines within a single institution, and on a global scale within these widely distributed communities. Increasingly one of the most important uses of information technology in research will be to support distributed collaboration. Collaborative technologies have garnered great interest, but are not yet in widespread use. Existing commodity tools are in some cases of dubious quality. Attention to the possibilities of this technology is often focused on the most elaborate technology, now useful only in demonstration or by the most sophisticated users. Considered broadly, however, collaborative technology includes services that range from futuristic applications to a variety of applications that can be put to immediate use enhancing the everyday productivity of even relatively inexperienced IT users.
UITS will deploy selected collaborative technologies that are of obvious immediate value, trial other technologies that are ready for advanced users but not yet fit for wide use, and experiment in the development of new collaborative technologies, as described in Actions 27 and 28a, b, c, d, and e, below.
Research Collaboration: Web-based Data Exchange and Collaborative Use of Research Software (Actions 27 and 28a)
The Indiana University Information Technology Strategic Plan notes: "At present, the Web is the most dramatic example of collaborative technology." Typical Web-based services, however, provide only a basic approximation of the highly interactive activity that characterizes collaboration. Immediate expansion of Web-based collaboration must focus on the key opportunities it holds for researchers: the opportunity to publish (to the entire Web or to specified audiences) data from ongoing research, and the opportunity to use the Web as a data input tool. UITS will establish a Web-based collaborative environment for exchange of data. This environment will create a completely new and open-ended environment for researchers to develop and deploy tools that permit collaborative exchange of data (with flexible security features that provide settings for multiple levels of privacy). UITS will thus create an arena where researchers are able to pool, share, publish, enter,and interact with large data sources regardless of geographic location.
- Create a research-oriented Web server that will provide researchers with a more open development environment for the creation of Web-based collaborative software solutions, and to house data for research collaboration. Deploy on this server applications such as Oracle Application Server and SAS/Internet as well as other major Web development tools that will facilitate Web-based collaborative software solutions.
Collaboratory Development - Advanced VR (Action 27 and 28b)
Collaborative technology can be put to routine use right away or in the very near future in many areas. Cutting edge collaborative technologies, such as virtual reality and tele-immersion, may reshape quite soon the way scholarly collaborations are carried out. IU has already joined the ranks of a select group of research organizations exploring these next generation technologies. For example, IU participated in a notable early demonstration of a collaborative virtual environment called NICE15 at SuperComputing '97. IU has been involved in experimentation in fully immersive collaborative technologies in association with Argonne National Laboratory (ANL), the Electronic Visualization Laboratory (EVL) at the University of Illinois - Chicago, and NCSA. IU should continue with experimentation at the cutting edge, followed by the development of pilot projects, and then implementation of digital collaboratories. Within five years, such collaboratories should be used by the University community with at least the ease found today in videoconferencing facilities, but with profoundly greater impact on the intellectual pursuits of the IU scholarly community.
- Create and fill a position for a senior technical expert to
focus exclusively on VR-mediated collaboration and tele-immersion
technologies. (Completed 3Q1999)
- Continue to participate in research projects with NCSA, EVL,
- Continue to participate in e-commerce virtual shopping research
with researchers in the Kelley School of Business.
- Continue development of advanced visualization applications
and tools such as 3DIVE(3 Dimensional Interactive Volume Explorer)
and XMView (for molecular visualization).24,
- Plan a staged upgrade of current VR facilities, including evaluation of new technologies and consideration of the expanded availability of VR facilities Common File System (Action 27 and 28c)
Collaborative applications depend upon a shared, distributed file system. Today, the network connects desktops with high performance computing resources. In many ways, however, workstations remain isolated from each other because of separate file systems and applications that are not "network aware." A common, networked file system that spans the entire University network can ultimately be integrated to create new possibilities in collaboration, information exchange, and exploration of data that will fundamentally alter the intellectual pursuit of knowledge.
The variety of file systems at IU impedes seamless, transparent access to user data. Although systems such as the NFS(Network File System) server at IUB and Network Appliance at IUPUI are widely used, these systems permit only limited interoperability. Furthermore, a typical user can have home directories on more than half a dozen different file systems.
The goal of the Common File System initiative is to provide IU faculty, students, and staff universal access to a distributed file storage system. The technical and support infrastructure currently being developed for the Massive Data Storage project, specifically the DFS (Distributed File System) infrastructure, can be leveraged to provide a more coherent and integrated suite of central file services, and to provide a common interface and interoperability among file systems.
One key to development of better central file system services will be to establish a principal of "two campuses,one architecture" between IUPUI and IUB. IUPUI and IUB now have three similar, but subtly and frustratingly different file systems: Bookbags and Lockers, Network Appliance and NFS server, and Jewel and Shakespeare file systems. New deployments should be architecturally identical across campuses to capitalize on economies of scale.
- Implement the first phase of the Common File System Infrastructure to replace reliance on NFS services at IUB (HPNFS and other NFS servers) and IUPUI(Network Appliance).
- Establish AFS (Andrew File System) as a supported production service for those scholars who require it in order to participate in inter-institutional and international collaborative research.
- Establish, test, and evaluate environments for desktop/client services that make use of shared, distributed common file systems.
- Determine evolution paths for common file systems, including eventual replacement of NetWare-based services.
- Identify funding for the Common File System beyond that in various current equipment replacement funds. The consolidation of file systems will offer services and storage capacity far beyond those covered by such funds today.
Desktop Videoconferencing and Application Sharing (Actions 27 and 28d)
Digital Video Streaming for Collaboration(Actions 27 and 28e)
RAC will actively collaborate with the Telecommunications Division of UITS in their work to implement these Action Items, described in detail in the Telecommunications Implementation Plan.26
Simulation-based Laboratory Courses (Action 15)
ACTION 15. UITS should evaluate the opportunities to partner with faculty in the sciences to experiment with simulation-based laboratory courses, and should be alert to other possible partnerships for the enhancement of instruction through simulation and visualization.
Classes with hands-on laboratory exercises are essential in many settings, such as advanced undergraduate instruction in the sciences. In other situations, a computer-mediated simulation of a laboratory exercise may be either the best pedagogical option or the only feasible option.27 In a large introductory-level class given on campus, laboratory-mediated simulation may be the best option for hands-on experience, particularly if the choices are either a well-executed simulation exercise or a poorly executed and crowded exercise in the laboratory. Computer-mediated simulation may be the only feasible option for an introductory class delivered via the Internet.
Many resources that are freely available on the Web can be used in simulation-based laboratory exercises. Likewise,the ongoing activities of the Center for Teaching and Learning (CTL) and the Teaching and Learning Technology Laboratory (TLTL) can facilitate the local development of such simulation exercises. Such virtual laboratory experiments have been implemented at IU and other universities, but their sophistication and level of deployment have been far lower than current capabilities allow.
IU's overall strategy for distributed education is described in Charting a Course Toward an Indiana Virtual University: Distributed Education Plan.28 Action 15 will be pursued within the framework of this plan. The Associate Vice President for Distributed Education, along with staff from RAC and the Teaching and Learning Information Technology Division (TLIT) of UITS, will partner to identify additional sources of support for this activity. In addition, RAC and TLIT staff will identify and encourage the use of existing simulation resources available via the Web. The Teaching and Learning Technologies Laboratory in Bloomington and the Center for Teaching and Learning in Indianapolis will seek opportunities to partner with and support the local development of laboratory exercises, particularly those that IU is uniquely positioned to create through its own intellectual leadership.
- Identify and highlight existing activities that demonstrate
the value of this approach. The partnership between AVL and
the IUB Department of Astronomy is one such example. Another
outstanding example is the strong partnership established by
UITS with the Department of Geology to support their outreach
to Indiana schools through their virtual seismology laboratory.
- Identify and publicize existing resources for simulation-based laboratories that are already available on the Web.
Massive Data Storage Initiative
ACTION 32. The University should evaluate and acquire high-capacity storage systems, capable of managing very large data volumes from research instruments,remote sensors, and other data-gathering facilities.
ACTION 43. UITS should implement massive storage technology for storage of the University's institutional data, migrate tapes over time to the new environment, and integrate this technology with database management systems to support image, sound and video data types.
A related UITS initiative plans for a complete restructuring of the central file serving capabilities. (See Actions 27 and 28c above.)
Data storage is a fundamental IT service for University users, comparable to the provision of computing cycles and network bandwidth. Major progress has been made on the Massive Data Storage project, starting with initial hardware installation in December 1998, and leading to establishment of HPSS as a production service in June 1999. With support from many in the UITS organization, and with the help of colleagues at IBM and the HPSS consortium, this storage system was brought online in record time. A number of substantial tasks lie ahead for the Massive Data Storage project, including expanding the HPSS service from an initial group of experimental users to a larger population of academic researchers. Also ahead will be major test and development work involving HPSS, DFS (Distributed File System), and Web interfaces to the data storage system. DFS (Distributed File System) is a facility for secure, scalable,and easy sharing of files among distributed computing systems, and a high-speed interface between DFS and HPSS will greatly facilitate analysis of very large data sets. Efforts will involve:
- Ongoing operation of a production HPSS and DFS service.
- Continued testing and development of advanced storage technologies.
- Application testing, demonstration, and evaluation to promote and support the use of new storage services.
A related task is the creation of new and better central file storage systems. This activity is described above,under Actions 27 and 28c - Common File System.
- Establish DFS and the HPSS/DFS interface as production services.
(DFS/HPSS in production August 1999.)
- Expand capacity of disk and tape storage.
- Implement a Massive Data Storage service at IUPUI.
- Expand the user base to serve a larger number of faculty and
graduate students with research applications. Continue to support
the early adopters and hardy users.
- Conduct testing and evaluation to greatly expand the user
base for various applications: large sets of scientific data,
digital video and audio, real-time data capture (e.g., from
instruments or remote sensing devices), long-term data archiving,
- Develop and evaluate Web gateway solutions for distributed
access to central storage, storage policy enforcement, etc.
- Test and evaluate the integration of Massive Data Storage
with other central IT services, such as messaging, Web serving,
- Conduct research and development to integrate work at Indiana
University with national and international massive data storage
and other high performance computing and networking initiatives,
including emerging "data grid" applications and services.
- Explore collaborations between massive data storage and digital library initiatives in the areas of metadata and asset management for digital content, digital audio/video, digital media preservation and digital archives,etc.
Digital Library Research (Action59)
ACTION 59. The University should develop a program of digital library research, and engage in national initiatives, to address the issues of innovative user services, creation and management of digital collections, the federation of distributed digital libraries, and the design of digital library systems.
By addressing a number of fundamental information science and technology problems in the development of digital libraries, Indiana University is positioned to take a leadership role in this area. This action will provide one-time and base funding to conduct digital library research and development, building on its successes in digital audio, digital video, electronic texts, and other areas. This action is, in part, contingent upon success in receiving competitive grants for digital library research and development, as external funds must provide a share of the expense. IU will undertake modest research and demonstration projects, funded primarily or entirely with IT Strategic Plan resources. These projects will make small but important advances in digital libraries and establish and demonstrate the ability of IU to succeed in this area. The majority of funds will be reserved for use as matching resources for major grant-funded projects in digital libraries.
- Conduct small-scale demonstration projects: digital media
delivery across wide-area networks between US and Japan (to
include evaluation of streaming media playback and control over
long-distance networks), and possibly between US and UK (to
include usability testing at remote sites).
- Establish a small prototype digital publication (electronic
pre-print)service, serving a discipline-specific scholarly community
centered at IU but with substantial international participation.
This demonstration project will determine IT and library support
requirements for operating a digital pre-print service and serve
as a field test for metadata/interoperability standards. (see
- Leverage IT Strategic Plan resources as matching contributions for future digital library research funding proposals.
Digital Library Infrastructure (Action 60)
ACTION 60. The University should develop a digital library infrastructure that will provide a common technical and organizational base for new and ongoing digital library programs.
Critical to the long-term success of digital library efforts at IU will be a coordinated approach to building a digital library infrastructure that will support teaching, learning, and research. The three chief components to a digital library infrastructure include: a) digital technology (servers, storage, networks, software,etc.); b) digital content (creation, acquisition, licensing, and subscription, etc.); and c) human and organizational resources to support the use of scholarly material in electronic form (reference, collection development, cataloging, etc.). This action will help fund the first of these, digital technology, with the expectation that funding of digital content and organizational resources will be primarily the responsibility of the campuses, schools, departments, and the University Libraries.
- Document existing and proposed digital library applications,
including digital audio, video and image services,electronic
text services, electronic reserve services, etc.
- Specify the overall technical design of the digital library
infrastructure. Document requirements for server, storage, and
network technology to support digital library applications,
and develop one-year,three-year, and five-year budgets and implementation
- Identify staff requirements to support the above applications
and technologies. Adjust work assignments of existing staff.
Develop position description and hire new staff.
- Identify University commitments (libraries, campuses, departments,others) toward digital content acquisition and development.
Access to Electronic Information Resources (Action 61)
ACTION 61. The University Libraries, with UITS, should provide students, faculty and staff at all campuses with convenient and reliable access to a comprehensive and coordinated collection of electronic information resources, on the campuses and off.
As noted in the previous action,electronic resources (i.e., digital content) are primarily the responsibility of campuses, schools, departments, and the University Libraries. As such, the planning and priority setting for this action falls primarily on these units and outside of UITS, as does the requirement to fund acquisition of electronic resources.
This action will provide funding for a staff position to integrate the operation of diverse electronic library resources with the developing digital library infrastructure at IU. A high priority for this position will be to analyze content and user requirements in digital library projects and applications, develop user interface design specifications, coordinate usability testing, and implement design changes suggested by testing feedback. This position, working for UITS, would serve libraries at IUB, IUPUI, and the regional campuses. The success of this arrangement will depend in part on the degree of coordination that exists among libraries and campuses in the licensing and delivery of electronic information resources.
- Identify the role and function of technical staff to coordinate
integration of electronic information resource services with
central IT services.
- Develop a position description and hire new staff. (Interface Design and Content Analyst position filled May 2000.)
Electronic Reserve Service (Action 62)
ACTION 62. The University should develop within its digital library program an "electronic reserve" service so that faculty can assemble and make available content in all media and formats: text, image, audio, or video; published or unpublished; digitized representation or original digital artifact; etc.
Access to course reserve materials(articles, book chapters, faculty notes, etc.) is an essential library service. Advances in distributed education create new demands for the electronic delivery of library materials and services. Digital library technology permits this service to be delivered to students at home, at work, and in the residence centers, or to students enrolled in distributed education courses. The IUPUI library has implemented a successful electronic reserve system. This action will fund implementation of a similar system at IUB, and provide staff to coordinate the operation of electronic reserve systems with other campuses as needed.
- Analyze,specify, and document functional requirements and
technical requirements for an electronic library reserve system
to meet faculty and student needs at IUB. Conduct similar analyses
of functional and technical requirements at regional campuses.
- Review the electronic library reserve system at IUPUI for
applicability to IUB and regional campuses. Review electronic
reserve systems in use at other universities, both custom-built
and those from commercial vendors.
- Determine and specify how electronic reserve system(s) will make use of the digital library infrastructure (Action 60). Specify how electronic reserves will be integrated with the library automation system, and how electronic reserves will be integrated with Oncourse or other Web-based course information and online learning environments.
Life-cycle Funding for Existing Digital Library Projects (Action 63)
ACTION 63. The University should establish sound funding for existing digital library initiatives (including VARIATIONS, LETRS, IMDS, others), and should provide support for other digital library projects of merit that are advanced in the years ahead.
IU has implemented a number of leading-edge digital library applications, but has supported them primarily with one-time funds. This action will establish a life-cycle funding plan for the ongoing operation and long-term success of these services.
- Assess information technology requirements for current digital
library applications and plan a multi-year schedule for ongoing
upgrade and life-cycle replacement.
- Determine priorities for equipment upgrade or replacement;
specify equipment and schedule installation.
- Establish an annual plan for equipment upgrade and replacement.
Electronic Records (Action 64)
ACTION 64. UITS, in partnership with the University Archives, Internal Audit, the Committee of Data Stewards, and others should develop a program to assure preservation of electronic institutional records.
UITS will work with the University Archives, Internal Audit, and others to establish retention and disposition schedules for the University's institutional databases and electronic records, and evaluate the liabilities involved in electronic record retention. This action will fund permanent staff to coordinate and support these analysis activities, and will fund project staff to help assess the University's current situation and develop a plan to correct deficiencies. The work of this staff will be allied with the University Information Services Division (UIS) of UITS, and will serve information needs University-wide. This action should work in tandem with, and help to support, the recommendations made in Action 40 (Data Administration), which are under the direction of the IU Information Technology Policy Office (ITPO).
- Define the scope of work and develop a work plan for an electronic
records management project.
- Identify staffing requirements; determine existing staff who
can be assigned and additional staff needed for the activities
- Depending on anticipated funding, hire new staff or temporarily
reassign current staff.
- Begin developing retention and disposition schedules for electronic records and institutional databases. Begin evaluation of institutional liabilities related to electronic record management.
Digital Media Preservation (Action 65)
ACTION 65. UITS, in partnership with the University Libraries, University Archives, and others should evaluate technologies and propose methods and standards to protect digital materials against media deterioration and technological obsolescence.
Increased use of digital media and increased reliance on digital content raise concerns about whether information stored digitally will be available when needed at some time far in the future. An issue of great concern is whether the digital media on which information is stored will deteriorate over time or be made obsolete by changes in technology. This action will fund technology and staff for an ongoing program of digital content preservation, to be conducted in alliance with the University's libraries and the UIS division of UITS. Planning for this migration will be undertaken jointly by UITS, the University Libraries, and the University Archives.
- Review research literature, research and development projects,
and other initiatives in digital media preservation.
- Define the scope of work, establish goals, and develop a work
plan for digital media preservation project.
- Identify staffing requirements; determine existing staff who
can be assigned and additional staff needed for the activities
- Determine information technology requirements, based on project scope,goals, and work plan.
Copyright Management, Intellectual Property, Authentication (Action 68)
ACTION 68. UITS should collaborate with the Copyright Management Center on developing policies and programs that advance the use of information technology and information resources, especially in areas of teaching and research, while limiting the University's liability exposure regarding intellectual property rights.
One of the most critical issues facing the broad deployment of networked information resources, not only here at IU but worldwide, is the potential barrier raised by concern for intellectual property rights. The University is fortunate to have the support of the Copyright Management Center (CMC), located on the IUPUI campus. As part of its mission, the CMC is "pursuing novel policies and programs that illuminate the complex interrelationships of intellectual property laws and emerging technologies." This action will fund a liaison function between UITS and the CMC to help implement and manage the technology needed to meet institutional responsibilities for intellectual property protection. This may include implementation of authentication and authorization systems, or other actions mandated by the "Digital Millennium Copyright Act" enacted by the 105th Congress in 1998.
- Define the scope of work and establish goals for information
technology in support of intellectual property and copyright
- Evaluate priorities (e.g., authentication services, access
management restrictions for "fair use" exceptions, etc.), and
establish a workplan for this initiative.
- Identify staffing requirements; determine existing staff who
can be assigned and additional staff needed for the activities
- Depending on anticipated funding, hire new staff or temporarily reassign current staff.
Associate Vice President Christopher Peebles leads the UITS Quality Initiative,29 an ongoing effort to enhance the quality of UITS services while improving their cost effectiveness. A significant part of the operational responsibility for the UITS Quality Initiative is rooted in the RAC division. In particular, RAC is responsible for the annual UITS customer satisfaction survey, which provides sound statistics on usage and customer opinions of UITS services. Full details on the results of the many years of UITS User Survey data and more information on survey methodology can be found at http://about.uits.iu.edu/divisions/rac/usersurv.html. RAC will be vigilant in all its activities in assessing user-perceived quality of services, meaningful usage metrics,30 and accurate product and process costs as established in the balanced scorecard methodology.31 This will in turn assure that RAC is delivering services viewed as valuable by its customers, providing resources that add demonstrable value to the productivity of the University, and is doing so in a cost-effective manner.
|Table 1. Quality Baseline for RAC Services|
|Quality Score||Percentage Use||Quality Score||Percentage Use|
|Overall Satisfaction With RAC Services*||3.8||47.5%||3.8||13.8%|
|Central Research Computing **||3.8||9.9%||3.7||7.4%|
|High Performance Computing**||4.0||5.1%||N/A|
|Central Instructional Computing*||3.7||16.2%||N/A|
|Unix Workstation Group**||3.9||8.6%||4.1||4.1%|
| * Faculty, Graduate, and Undergraduates surveyed
** Faculty and Graduate Students surveyed
UITS in general and RAC in particular have a long history of careful attention to the results of the UITS User Survey. The 1998 survey results regarding RAC services are presented in Table 1. Ratings are done on a five-point scale, with a one (1) indicating "Very Poor" and a five (5) indicating "Excellent." The table displays the average score for each service, the percentage of users of each service who are satisfied with it, and the percentage of survey recipients who use each service. Of these three items, only the satisfaction percentage requires further explanation. This figure is the percentage of survey respondents who use a particular service and who rated that service a 3(Good), 4 (Very Good), or 5 (Excellent). Since the UITS customer satisfaction survey is sent to a stratified random sample of the Indiana University community at IUPUI and IUB, it provides a representative view of the average perception of UITS among its customers on these campuses.
Overall, approximately 47.5 percent of the Bloomington and 13.8 percent of the Indianapolis faculty and students use one or more RAC services. The overall rating for RAC services averaged a 3.8, which equates to a score very close to "Very Good" on the five-point scale. Ninety-five percent of the users on the Bloomington campus and ninety-two percent of the users on the Indianapolis campus are at least satisfied with RAC services overall.
A more detailed analysis of the survey results for each service group with RAC follows:
High Performance Computing. Users of UITS HPC systems give this service a solid "Very Good" rating (an average score of 4.0 and a satisfaction rating of 91%),indicating a high degree of satisfaction with UITS supercomputers. (For the 1998 survey, these were the IBM RS/6000 SP and the SGI Origin 2000). This service received a 100% satisfaction rating for the two years prior to 1998. It is important that RAC try to return to the previous 100% satisfaction rate. Given the size of the user community and the amount of staff (recently increased) devoted to the support of these systems, it is an area where RAC can realistically hope to achieve 100% satisfaction levels repeatedly.
Central Research Computing. The Steel cluster and the provision of serial research applications on the IBM RS/6000 SP receives "Very Good" ratings from users on the Bloomington and Indianapolis campuses. Here too, it would be useful to see what it would take to move users and their ratings from "Very Good" to "Excellent" and from 3.0s and 4.0s to perfect 5.0s.
Central Instructional Computing Services. These services are currently offered on the Bloomington campus by the RAC division and on the Indianapolis campus by the University Information Systems division. These services also score in the "Very Good" range, but clearly there is room for improvement. The implementation of Actions 27, 28, 30, and 33 should lead to improved user satisfaction with these services.
Center for Statistical and Mathematical Computing. The Stat/Math Center received an exceptionally high rating among Bloomington users - 4.1 and 99 percent satisfied -and a lower score among Indianapolis customers - 3.8 and 83 percent satisfied. The difference can be accounted for by the amount of time Stat/Math personnel could spend on the Indianapolis campus in 1999. With the assignment of full-time research applications support specialists in Indianapolis in late 1999 and 2000, user satisfaction should increase markedly.
The Unix Workstation Support Group(UWSG). IUPUI users rated UWSG very high - 4.1 and 95 percent satisfied - attributable largely to the assignment of full-time personnel there in 1999. IUB users, on the other hand, rated UWSG a 3.9 and 86 percent satisfied, a drop from 1998 ratings. The latter rating is due in large measure to the fact that UWSG in Bloomington was understrength throughout the year, as corporations and other divisions of UITS systematically recruited highly qualified and motivated persons from UWSG.
Library Electronic Text Resource Service(LETRS). LETRS received a "Very Good" rating, 3.9 and 88 percent satisfied, from the Bloomington faculty and graduate students who used their services. This service, which is a partnership of UITS and University Libraries, should benefit from the additional resources and expanded commitment of the Digital Library Program. At the moment, LETRS is decidedly overcommitted and understaffed. Almost 20 percent of Bloomington faculty and graduate students use this service (and there are also undergraduate users of their services), so the ratings for LETRS should be viewed in a very positive light.
RAC services, as perceived by its customers, currently fall roughly in the "Very Good" range. This evaluation comes as a result of years of investments in RAC hardware, software, and staff. Recent increases in investments have been made possible as a result of funding for IU's Information Technology Strategic Plan. Ongoing attention to the results of the UITS survey has also contributed to this generally positive view. Indeed, the historical results of the UITS survey show consistent improvements in customer satisfaction with research and academic computing services over the past several years. Still, approximately 10% of the users of RAC services find them unsatisfactory overall, and specific services have received unsatisfactory ratings by as many as 17% of their customers. This information can be viewed in two ways. One is to consider the results of the American Society for Quality's national benchmarking surveys, which find that 25 percent of customers rate the services they receive from service providers generally as inadequate. This view provides justification for what many members of the IU community have said anecdotally - the services RAC offers are better than comparable services at many other institutions. On the other hand, our ultimate goal is "six sigma" quality - three unhappy customers per million. Given the size of the IU community, this translates into a goal of all customers being at least satisfied with RAC services. RAC's current operational target is for a satisfaction rating of 95% for all services. RAC leadership considers anything less than that a problem in need of remediation. Once RAC achieves that target across the board, the operational target will move to 99% satisfaction. In this way, RAC will improve its services so that eventually all customers find them at least satisfactory, and so that most customers find them to be excellent.