Welcome

Dori Helms, Vice President for Academic Affairs and Provost
John Kelly, Vice President for Public Service and Agriculture
Chris Przirembel, Vice President for Research and Economic Development

Clemson Universities three “mission VPs” welcomed attendees and expressed their individual and collective support for cyberinfrastructure (CI) development and application, at the University and beyond. Vice President Kelly, who oversees 60 off-campus locations in the U.S. and abroad and travels frequently, anticipates that CI will support distance learning, seamless remote presence, and collaborative research in such disciplines as environmental systems management. Provost Helms, a biologist by training, envisions CI as both a systems integration challenge and a facilitator of increasingly interdisciplinary and collaborative teaching, learning, and research. Vice President Przirembel perceives CI as a major technology enabler for research and economic development for both the university and the state of South Carolina.

(20-minute video on iTunesU)

Keynote: Adventures in Cyberinfrastructure: Observations of an Accidental Tourist

Mark Lundstrom, Don and Carol Scifres Distinguished Professor of Electrical and Computer Engineering at Purdue University; founding director of the National Science Foundation’s Network for Computational Nanotechnology (NCN)

In a wide-ranging case study of the power and impact of cyberinfrastructure on one faculty member’s teaching and research in the field of nanoelectric devices and materials, Professor Lundstrom described fifteen years of collaborative research and software development work at Purdue. Quoting Richard Hamming, “The purpose of computing is insight – not numbers,” he provided four rationales for his investment in IT at Purdue: to develop understanding, to interpret experiments, to explore new devices, and to set the stage for more serious simulations.

Lundstrom then introduced a “computational science and engineering continuum,” with theorists and modelers (“closer to the problem”) at one end, algorithm developers and high-performance computing specialists (“closer to the solution”) at the other, with simulation and computer-assisted design specialists in the middle. An ongoing challenge, he stated, is to foster collaboration among these three groups.

PUNCH – Purdue University Network Computing Hubs – was an early and successful project that made advanced simulation capabilities available via a web browser; between 1994 and 2002 PUNCH recorded more than seven million hits. In 2002 a multi-university effort with Purdue in the lead, Network for Computational Nanotechnology (NCN), was launched with NSF funding “to pioneer development of nanotechnology from science to manufacturing through innovative theory, exploratory simulation, and novel cyberinfrastructure.”

NCN’s ambitious software strategy includes facilitating sharing of software tools emerging from research, disseminating high-quality simulation codes, developing specialized tools for experimentalists and educators, and promoting intelligent, critical use of simulation. A key component Rappture – Rapid Application Infrastructure – seeks to standardize interfaces, improve usability, provide a complete record of each simulation, facilitate software development, and enable ambitious workflows. Other components include a middleware system architecture named Maxwell’s Daemon, an extensive toolkit, education and training materials, and collaboration capabilities.

NCN now counts over 65,000 users all over the world, with 100,000 projected within the next year. A majority of users are graduate students, with undergrads, faculty, and nonacademic scientists and engineers accounting for 18%, 13%, and 13%, respectively. Usage statistics can be found at www.nanoHUB.org.

The underlying software structure supporting NCN is available for support of other disciplines through a project at Purdue called Hubzero. Other projects using the Hubzero technology focus on cancer care engineering, advanced manufacturing techniques, global engineering education, pharmaceutical product development and manufacturing, heat transfer, and clinical/translational research. For more information see hubzero.org.

Lundstrom then provided examples of the broad impact of this work. In one year over 10,000 users viewed online seminars and courses developed by Purdue Professor of Electrical and Computer Engineering Supriyo Datta. Lundstrom’s biennial course in Advanced CMOS Devices offered to 40 students at Purdue is taken by 4,000 remote students during the same period. A forthcoming article in Nature by M. Ashrad Alam, et al., references the NanoHub, which enables interested readers to explore simulations with alternate parameters. And the June, 2008 Review of Scientific Instrumentation highlights a NCN-enabled project, VEDA, Virtual Environment for Dynamic AFM (atomic force microscope).

Referencing a key tenet of the November 2006 NSF Strategic Plan, “The conduct of science and engineering is changing and evolving. This is due, in large part, to the expansion of networked cyberinfrastructure.” Lundstrom predicted that the benefits will accrue to those who “dive in and learn how to do it.” He pointed to the NanoHub as an example of a community of researchers who require access to expensive specialized equipment and data that are appropriate to share in what Foster has termed, “service-oriented science.”

Reviewing lessons learned from the NCN experience, Lundstrom suggested:
• It takes a dedicated team with a vision and something special to share
• Team needs people who are both “close to the problem” and “close to the solution”
• Professional IT talent is required
• Adaptation and evolution are also required
• CI is expensive, not only for software development but also for assessment
Yet NCN is a work in progress, still grappling with such issues as:
• Refining and expanding the software collection
• Evolving from a “resource” to a “community”
• Continuing to enhance the infrastructure
• Broadening coverage of the nanotechnology domain
• Growing the user base
• Sustaining the effort

Professor Lundstrom closed by predicting that research projects and universities with a sound CI strategy will have a strong competitive advantage, and exhorting the audience to explore the broader applicability of his experience developing and exploiting cyberinfrastructure to create an international collaborative research and educational community.

(58-minute video on iTunes U, plus PowerPoint)

What is Cyberinfrastructure?

Russ Hobby, Cyberinfrastructure Program Architect, University of California, Davis, on assignment as Program Manager, Internet2

Mr. Hobby observed that there are numerous definitions of Cyberinfrastructure (CI), a term attributed to the National Science Foundation¹ that has been adopted by other Federal agencies interested in promoting collaborative research. Federal interest in and support for CI is increasingly cognizant that much modern research is carried out via collaboration of researchers from multiple institutions forming a “virtual organization.”

Consider, for example, “Jane,” an environmental researcher collecting data via remote sensors, transferring the data to a storage medium, analyzing and visualizing the data, using them to develop and refine a model, perhaps via simulation, and publishing the results. Some of these processes may also be performed using specialized, remote resources, and some may be done locally, perhaps with remote collaborators. Access control and security must be maintained, ideally without imposing undue complexity to the collaborators, a robust network is required, and data management and archiving expertise will likely come into play. Jane and her research collaborators will call on the skills of numerous campus and remote IT and library experts, functioning, ideally, to provide a coherent and seamless environment in support of their collaboration – in short, a cyberinfrastructure.

CI is no longer just for scientific research, however. CI and its components – networks; access control, security, and identity management services; computational and storage resources; visualization; data management and archiving; collaboration tools – are increasingly valuable for research in other disciplines, for instruction, university administration, and outreach. And, as more users become dependent on CI for more and different activities, additional investment is required to “pull it all together” in pursuit of a unified, high-performance, and cost-effective collaborative environment.

“CI Days” workshops are one approach to enhancing understanding of the requirements for CI and to coordinate functions and roles locally, regionally, nationally, and internationally. Sponsors of CI Days are Educause, Internet2, National LambdaRail (NLR), Open Science Grid, TeraGrid, and the NSF International Research Network Connections (IRNC) program. Workshops have been held for individual campuses as well as for regional clusters of institutions. For more information, see http://cidays.org/.

In conclusion, Hobby opined that CI is moving to include all aspects of information technology as a coordinated system. While NSF has provided a vision, much work is needed to define an architecture and develop an implementation plan to integrate computation and storage resources, data repositories, and tools to facilitate analysis, visualization, simulation, and collaboration.

(23-minute video on iTunesU, plus PowerPoint)

Cyberinfrastructure and the Networked Information Community

Clifford Lynch, Executive Director, Coalition for Networked Information (CNI)

Mr. Lynch noted that CNI is a non-profit association of some 200 member universities and related organizations, including Clemson, “working a broad agenda” around digital content and technology in support of research, teaching, and learning; see www.cni.org.

According to Lynch we are seeing profound and extensive changes in the process of scholarship, driven by information technology, including advances in sensor systems, data management, data mining, visualization, and simulation. Impacts are most evident today in sciences and engineering; however, changes are rippling through social sciences, quantitative and non-quantitative, and humanities.

Lynch defines cyberinfrastructure as a way of talking about not only technical systems but also human systems and policy structures that support and facilitate these changes in scholarly practice. CI is not an end in itself, nor is it restricted in application to research or to higher education; applications are emerging in support of instruction as well as outreach to the K12 community². He then discussed three broad challenges posed by CI:

• The nature of CI and issues facing campuses. Early leadership for CI has come at the national level from science funding agencies – not only in the U.S., but also in other countries -- often focused on major, well-supported projects, such as the Hubble Telescope and Large Hadron Collider³. We are now seeing CI applied on medium-sized and small projects that require significant campus-based support. Some campus challenges are non-technical – including innovative support services and policies that provide incentives for resource sharing.

• Data management, curation, and reuse. Perhaps the greatest challenge will be data – either observational or derived data from analysis and simulations – that need to be managed and preserved. Funding agencies increasingly are attentive to these issues, and some are mandating plans for data sharing when reuse of data assets makes sense. Publications are also changing form, with some little more than de facto “stubs” for datasets that are made available in repositories. Another challenge is management of data in non-digital form, including artifacts, specimens, photographs, sound recordings, et al⁴. Will long-lived datasets be discipline based, institution based, or both? How will repositories be funded, since funding agencies tend to focus on the research only? Is this responsibility a natural extension of institutional research libraries? New collaborations will be required among librarians, IT experts, information scientists, and discipline specialists.

• Collaboration and virtual organizations. Some important concepts predate CI, per se, such as the “collaboratory”— a “center without walls, in which the nation’s researchers can perform their research without regard to geographical location, “ envisioned by William Wulf of NSF in the late 1980s⁵. Today we speak of “virtual organizations” that are required to address collaborative challenges, but could and should disappear when the collaboration is concluded. How will shared assets be managed during the active phase, and who will be responsible for these assets over the longer term?

Lynch closed his remarks by urging the audience to think beyond hardware and software to new organizations, new organizational collaborations, and new specialties within information science and information technologies that will be needed for us to have a successful cyberinfrastructure.

(25-minute video on iTunesU)

Panel: Condo, Condor and the Grid

Sebastien Goasguen, Assistant Professor, School of Computing (co-chair)
Jim Pepin, Chief Technology Officer (co-chair)

Emil Alexov, Associate Professor of Physics
Steve Stuart, Associate Professor of Physical Chemistry
Bob Latour, Professor of Bioengineering
Dan Noneaker, Associate Professor of Electrical and Computer Engineering

Professor Goasguen introduced the session with a high-level, five-stage overview of CI history, starting with high-performance computing initiatives (such as the NSF-funded supercomputing centers in the 1980s), leading to grid computing in the 1990s, then major “E-science” initiatives, followed by a “long tail” phenomenon as many new research and educational users come aboard, and, finally, broader, social interaction-focused participation, exemplified by today’s FaceBook and SecondLife.

Using a modern residence as metaphor, he displayed a diagram of a “CI house,” with (a) a set of resources (computers, televisions, stereos, PDAs, gaming devices, cameras, phones, et al.) connected by wired and/or wireless networks; (b) a set of applications and interfaces (the web, email, telephony, calendars, repositories of photos, music, and videos, and such); and (c) a unifying structure, often referred to as “middleware” that, ideally, links and organizes the resources and applications. (Of course in most homes today, the constellation of devices does not deliver a satisfactory and unified set of services, due to immaturity of middleware and other problems, including conflicting proprietary approaches – Apple versus Microsoft, phone company versus cable company, et al.)

At the university level, the same three “stacks” can be found, although the scale and complexity are substantially different. Interestingly many of the applications and interfaces at the “top” layer are similar or identical, including web browsers, email, and telephony. The resources at the bottom layer, however, are often massive and heterogeneous, leading to even greater challenges for the middleware layer – which users are authorized to access specific resources, for example.

Turning to the topic of the session, Goasguen explained that Condor, Condo, and Grid are different approaches to providing bottom-layer resources:

• Condor: a system to assemble unused desktop campus resources to solve large problems
• Condo: a paradigm to manage dedicated computing resources with multiple ownership
• Grid: a system of national resources for those whose needs can’t be met locally

Professors Alexov and Stuart are major users of the Clemson Condor system⁶. Professors Latour and Noneaker take advantage of the Clemson Condo cluster, named Palmetto. Goasguen noted that one faculty member, Economist Paul Wilson, currently uses national grid computing resources, but he was unable to participate in the CI Days event.

Professor Alexov’s Computational Biophysics and Bioinformatics Lab is focused on predicting protein-protein interactions and the corresponding 3D structures of protein-protein complexes by homology and using the 3D structures to compute biophysical characteristics (www.ces.clemson.edu/compbio). This research involves calculating millions of cases, using non-trivial linear and non-linear mathematical functions, and Condor provides the computing power.

Professor Stuart researches chemical sputtering in nuclear fusion, a component of a $20 billion international project that aims to demonstrate the scientific and technical feasibility of fusion power (www.iter.org). In particular his team seeks to predict energetic impacts on the wall of the reactor via simulations of thousands of atoms, millions of atomic interactions, ten thousand to one million timesteps, and 10,000 samples. Collectively the computing power required can be measured in “CPU-decades,” so the team makes use of Condor, which provides an average of 1,400 processors, resulting in an average speed of 730 MFLOPS. (The work also leverages the Condo cluster, and supercomputers at Oak Ridge National Lab.)

Clemson CTO Jim Pepin introduced the Condo cluster and its rationales: electrical and cooling requirements of advanced clusters that stress the infrastructure of most academic buildings, technical complexity of cluster software management that distracts scientific teams, and the prospect of creating a higher-performance cluster with massive storage capacity by assembling resources from multiple grant-funded projects, augmented by investment from the University. Unlike Condor, which is ideally suited to highly-parallel (i.e., essentially disconnected) computations, Condo is tailored to problems requiring high-speed interconnection among multiple processors and storage (https://citi.clemson.edu/hpc).

At Clemson funded researchers purchase computing nodes, and the University provides racks, networking, storage, power, cooling, and management, plus additional nodes. “Owners” get first priority on nodes they purchase, while University nodes form a pool that is allocated by a faculty committee. The result is, currently, a 512-node cluster, with 4,000 cores, resulting in 30 TFLOPS of processing power, architected to be compatible with supercomputing facilities at national centers, in the event that Clemson researchers need even more computing power. The Clemson cluster ranked 62nd in the June 2008 semi-annual list of worldwide supercomputing sites, and fifth among non-federally-funded U.S. universities (http://www.top500.org/).

Professor Noneaker described his work in wireless communications research, including wireless communication systems, mobile ad hoc networks, and applied electromagnetics. Since January 2007, when his research group depended on a small, self-administered Linux cluster, to spring 2008, with a 12.5% ownership stake in the Palmetto Condo cluster, computing power has increased 15 fold, and total computing power to which they have access has increased by a factor of 100. Their work employs Monte Carlo simulations as well as numerical analysis, requiring hundreds of independent jobs running concurrently. The increased computing power enables the team to examine more design alternatives, to evaluate designs over a wider range of parameter values, and to employ higher-fidelity models of the link environments.

Professor Latour uses simulation to study implants and bodies’ reaction to them, including molecular-level modeling of implant surface biocompatibility. This research was originally performed on a Sun workstation, which was replaced by a 25-node private cluster that proved difficult to support. Professor Latour is another satisfied investor in and user of the Palmetto cluster.

(50-minute video on iTunesU – missing Latour -- PowerPoint available for Alexov and Noneaker)


Panel: Educational Dimensions of Cyberinfrastructure

Barbara Weaver, Manager, Instructional Service; Chair, Teaching with Technology Community (co-chair)
Jeff Appling, Associate Dean of Undergraduate Studies (co-chair)

Roy Pargas, Associate Professor of Computer Science; developer of MessageGrid
Brian Bolt, Instructor of Animal and Veterinary Sciences
Cynthia Haynes, Director, First-Year Composition and Associate Professor of English
Chris Peters, Associate Professor of Education; Director, South Carolina Center of Excellence for Instructional Technology
Cindy Pury, Associate Professor of Psychology; Faculty Editor, Community of Undergraduate Journals Online (CUJO)
Emily Gore, Librarian for Digital Initiatives, Cooper Library

Barbara Weaver introduced four “p” themes to be addressed by the panel: the role of play in education, student performance, software and hardware platforms that can enhance education, and the growing importance of hardware and software portability.

Jeff Appling referenced the “Techno-CroMagnon Theory” of Mark David Milliron, “’technology = good.’ The basic assumption was that the technology would almost magically drive business, transform medicine, and improve learning.” ⁷ Appling reported that many educators at Clemson and elsewhere have moved well past this “theory,” and are exploring more sophisticated approaches to teaching, learning – and assessment. For example, Appling himself no longer lectures, but instead teaches a more interactive course based on MessageGrid, a web-based software tool, developed at Clemson under the direction of Professor Pargas, to facilitate recitation and interaction in a laptop-enhanced classroom.

Pargas described Message Grid as a system that subsumes student submission of assignments; “clickers” for in-class assessment and polling; support for text, video, audio, and GPS; an “ink” function for input of diagrams and equations; integration of laptops, PDAs, and smart phones; and support for rubrics to facilitate evaluation of online discussions.

The panel was asked to speculate on what the future holds for education and technology. Pargas opined that performance assessment will improve, notably an enhanced capability for instructors, in a timely fashion, to ascertain which students are doing well and which need intervention. Professor Haynes, who deploys the massively multiplayer online role-playing game, “World of Warcraft” (http://www.worldofwarcraft.com) in her teaching, also uses role-playing in her research. In addition to more sophisticated gaming environments, she anticipates more portability, not only in multimedia devices but also in open-source (non-proprietary) software.

Emily Gore believes the library remains the center of the increasingly-digital life at Clemson; there is even a Facebook group named “Cooper groupies.” Under development are an “E-Research Portal” that will provide access to a wide range of scholarly resources as well as extension of the library’s traditional expertise and support services to unpublished materials and research data. Brian Bolt noted that instruction in animal and veterinary sciences requires “field” data collection and analysis; laptops are mandated today, and PDAs and smartphones could replace them in the future.

Pargas noted that Clemson benefits from an interdisciplinary group of faculty and IT support staff, the Teaching with Technology Community, which has been meeting weekly for nearly six years to share experience and insights about pedagogy and technology. The group’s Fall 2007 Teaching with Technology Symposium was highlighted by the dedication of a “sandbox classroom” in space that once housed an indoor swimming pool (http://www.clemson.edu/newsroom/articles/top-stories/sandboxclassroom.php5).

Professor Prury described her work with Community of Undergraduate Journals Online (cujo.clemson.edu). CUJO introduces students to the processes of scholarly writing, peer review, editing, and publication, and, thereby, prepares them to become publishing scholars. Ten CUJO journals are currently online, covering such disciplines as psychology, criminal justice, management, and teacher education.

Professor Peters, whose academic specialty is instructional technology was asked how cyberinfrastructure could help create future K12 education. Peters noted several innovative K12 projects, including Digital Xpress, a web-based collaborative student writing environment (mydigalxpress.com) and Club245, an after-school program to assist children who need help with school, such as tutorial assistance and life skills (itcenter.clemson.edu/club245). He expressed skepticism of educational master planning, per se, yet suggested that K12 education faces many of the same challenges as post-secondary education. He warned that distance learning technology now enables teachers to “bore and confuse students farther away than was possible in previous generations” and that new technology does not automatically engender effective pedagogy. That said, he suggested that educators explore the potential of social networking, notably the capability to create specialized social networks offered at the site, www.ning.com.

Regarding the future of reading, Haynes noted that some faculty colleagues reported difficulty reading long email messages, while students engaged in game playing may read energetically. Pargas observed that while students in his computer science classes don’t read the textbook on data structures, he believes there are better, interactive, methods for learning such topics.

An audience member asked about the impact of CI on informal education, workforce development, and economic development. Peters suggested that CI facilitates access anytime, anywhere to the best educational materials for those who are motivated, while unprepared or unmotivated students may exit such programs prematurely.

As for assessment, Haynes observed that while CI provides new opportunities for data capture in teaching and learning processes, the challenge remains to design and validate appropriate assessment schemes.

Regarding unequal access to educational technology, Peters acknowledged that CI developments are often expensive, and that successful technology-enabled teaching can exacerbate the so-called digital divide in K12. One high-profile initiative to address this concern is One Laptop Per Child, originally targeted at developing countries but likely to be deployed in some form in the U.S. as well (laptop.org).

The final audience question concerned support for CI in university libraries. Ms. Gore observed that Clemson and other libraries are undergoing reorganization and hiring staff with new skills to respond to these new challenges and opportunities.


Panel: Cyberinfrastructure in support of Architecture, Arts and Humanities; Focus on Access to Collections, Collaborations and Tools

Clifton S. M. “Chip” Egan, Interim Dean, College of Architecture, Arts and Humanities; Alumni Distinguished Professor Emeritus of Theatre (chair)

Dina Battisto, Associate Professor of Architecture
Jan Rune Holmevik, Assistant Professor of English and co-chair of the Rhetorics, Communication and Information Design (RCID) Colloquium on Serious Games
Bruce Whisler, Director of Music Technology and Assistant Professor of Music

Dean Egan, a lifelong aviation enthusiast, likened CI Days at Clemson to a meeting of the Experimental Aircraft Association, which attracts equipment suppliers, aircraft designers, pilots, hobbyists, and the general public, who enjoy the daily air shows. Egan likened the panelists to pilots, who use the tools and have ideas about how the tools can be improved.

Professor Battisto addressed the increasingly paperless study and practice of Architecture and the concomitant demand for robust IT solutions. At Clemson, Architecture is taught on the main campus as well as remote campuses in Charleston, Barcelona, and Genoa, so pedagogy is dependent on distance education, remote collaboration, and virtual studio capabilities. The program makes use also of 3D design software and fabrication devices as well as sophisticated simulation and rendering systems, which are computationally intensive. Battisto closed her presentation by demonstrating a 3D animated fly-over of a major tropical resort project on the paperless “drawing board” of a firm led by a Clemson Architecture graduate.

Professor Holmevik observed that the term “cyberspace” was coined by William Gibson, author of Neuromancer (1984) and a NASA spokesman credited science fiction literature, the Star Trek television series, and Star Wars films with inspiring some international space station developments. How then, can today’s humanities faculty inspire students to imagine our technological future?

Holmevik teaches a course on Visual Communication for which students design multimedia projects. Recent student projects include “digital remix” commentary on national politics, a film exploring family racial conflicts, and an exploration of tattoo art, using Second Life. Such projects require new and sophisticated technologies and support services heretofore unknown in the typical English department. Students in the course also write a thirty-page term paper and thirty pages of blog commentary, so traditional composition skills remain important.

Professor Whisler spoke of the need for managed digital storage of audio materials as a response to the challenge of preserving obsolete media. Recorded sound, a unique form of historical documentation, has existed for only about a century. Countless historical items in a variety of formats are endangered, however, including tape, film, wax cylinders, records, and associated printed materials. In addition to physical deterioration of media, threats include obsolescence of playback equipment and lack of experienced engineers.

The preferred preservation strategy is to convert the historical content to digital form based on optimal playback of the original analog items, using a 96 KHz sampling rate and 24-bit word size. This high-resolution digital capture results in 30 MB of data per minute, about three times the storage requirements of current audio CDs. The recommended archival format is a Broadcast Wave File (.bwf), stored on hard drives in two separate locations. Estimates of the data storage requirement in this field range up to 50 PetaBytes in the U.S. and 350 PetaBytes, worldwide (excluding redundancy).

The National Endowment for the Humanities is awarding grants for digital archiving projects, with the Sound Directions Project at Harvard and Indiana Universities as a prominent example (www.dlib.indiana.edu/projects/sounddirections/). Whisler concluded by calling for more collaborations, between IT and library experts and among universities, to address this archival crisis.

(56-minute video on ItunesU; PowerPoint and related materials for Battisto and Whisler)

Panel: National Perspective – Feedback and Advice to Clemson

Scott Lathrop, TeraGrid Director of Education, Outreach and Training (EOT) at the University of Chicago/Argonne National Laboratory (chair)

Ann Doyle, Director of Arts and Humanities Initiatives, Internet2
Patrick Dreher, Director of Advanced Computing Infrastructure and Systems, Renaissance Computing Institute (RENCI), Chapel Hill, NC; Research Scientist, Laboratory for Nuclear Science, Massachusetts Institute of Technology; Co-chair, Educause Net@edu Campus Cyberinfrastructure Working Group
Susan Fratkin, Principal, Fratkin Associates, Washington, DC; Consultant to Educause, Internet2, and Southeastern Universities Research Association (SURA); Washington liaison for the Coalition for Academic Scientific Computation (CASC)
Ronald R. “Ron” Hutchins, Associate Vice Provost for Research and Technology and Chief Technology Officer, Office of Information Technology, Georgia Institute of Technology; Vice chair, Southern Light Rail (SLR); Board member, National LambdaRail (NLR); Technical lead, Southern Crossroads (SoX)
John McGee, Manager, Cyberinfrastructure Development, Renaissance Computing Institute (RENCI), Chapel Hill, NC; Engagement Coordinator for the Open Science Grid

Sue Fratkin provided information about Federal support for Cyberinfrastructure as well as the CI activities of the Southern Universities Research Association (SURA). Primary funding agencies for high-performance computing include HSF and the Departments of Defense and Energy, while both NSF and the National Institutes of Health fund CI, per se. Major reports of the President’s Information Technology Advisory Committee (PITAC) and the President’s Council of Advisors on Science and Technology (PCAST) have helped to create a national vision for CI. Key messages of these reports include the requirement for long-term investment, additional research in both hardware and software, and implications for workforce development. By 2007 federal funding for high-performance computing totaled $3.1B, up from $366M in 2000. Funding for CI totaled $385M in FY’07 and ’08, with prospects for ’09 difficult to predict.

Fratkin observed that the Coalition for Academic Scientific Computation (CASC) now comprises 54 centers, including Clemson. Some of these centers receive federal support, and others receive state funding, including Louisiana, North Carolina, Ohio, and South Carolina. Some of the same institutions, including Clemson, are active in SURAgrid, a collaborative project to enable CI-supported research and education through grid computing. SURA has negotiated discounts on HPC systems with both Dell and IBM.

Patrick Drayer described CI-related activities of Educause, a consortium of some 2,200 universities and colleges focused on advancing higher education by promoting intelligent use of IT. Drayer co-chairs a Campus Cyberinfrastructure working group to assist universities in developing strategies, plans, and applications of CI. Educause’s Center for Applied Research (ECAR) has recently published reports on CI-related topics, and Educause is one of the sponsors of CI Days.

Dreher emphasized that all institutions face the challenge of sustaining CI initiatives, owing to the rapid obsolescence of advanced technology, the rapid growth of data repositories, and the difficulty of attracting and retaining the required expertise.

Ronald Hutchins applauded Clemson’s success in assembling for CI Days faculty and students from multiple disciplines, together with technical experts from the university and elsewhere. Sharing Professor Holmevik’s admiration for Neuromancer, Hutchins assigned the novel by the visionary but non-technical author for his Georgia Tech class in data communications. Given the rapid advances in computing, communications, and storage technologies, as well as higher education’s collaborative efforts at the state, regional, and national levels (for example, NCLR and SCLR; SLR and SoX; NLR and Internet2, respectively), he urged that the audience take the opportunity to think visionary thoughts and let CI be built in response to the diverse needs of science and the arts, high performance and disaster recovery.

Ann Doyle of Internet2 agreed with Hutchins that collaborations between artists and technologists can be mutually rewarding, not only because of the novel ideas the artists and performers generate, but also owing to the advanced capabilities they often require. She noted that Clemson is one of 209 university members of Internet2, together with many affiliate members (including libraries, museums, and orchestras), international network partners, corporate partners, and K12 participants. Arts and Humanities resources available include the Indiana University Digital Music Library, the Shoah Foundation Institute for Visual History and Education containing over 32,000 Holocaust interviews, over 2,800 programs organized by the Research Channel, and student-produced programming of the Open Student Television Network (OSTN). Other innovative network applications include remote delivery of music master classes, multi-site access to live performances of orchestras, dance and theatre companies, and exploration of education gaming and virtual worlds.

John McGee described the Open Science Grid (OSG) as a framework for large-scale resource sharing in the High-Energy Physics community, which addresses not only technological but also policy and social requirements for sharing. He argued that campus and inter-university resource sharing is both necessary and practical, offering new levels of scale and new modalities of collaboration compared to independently owned and managed facilities. McGee described the “Engage Virtual Organization,” created to support scientists wishing to use OSG via embedded immersive engagement (www.eie4ci.org). Projects being supported include massive Monte Carlo simulations for protein folding and natural language processing in Information/Library Science. He closed by saluting the CI initiatives and expertise at Clemson.

Scott Lathrop introduced TeraGrid another virtual organization in which eleven resource providers seek to operate as one organization. Currently, thanks to recently funded supercomputing centers at UT Austin and Oak Ridge National Laboratory, supply – nearly a PetaFLOPS of computing power -- exceeds demand, and more facilities are scheduled to come online soon. TeraGrid also offers a centralized help desk, remote visualization services, over a hundred scientific data collections, and distributed technical support. Tech support ranges from assistance on parallelizing software code to optimizing work flow across the distributed resources.

TeraGrid resources are provided free to academic researchers and educators. Allocations are made available at three levels: start-up (up to 30,000 processor-hours); medium (up to 500,000 hours); and large (over 500,000 hours). Currently roughly 1,000 principle investigators use TeraGrid resources, with more to be served via connections to Science Gateways (such as NanoHUB). TeraGrid offers an extensive set of training, professional development, and curriculum development resources, many published through the Computational Science Education Reference Desk (CSERD) project of the National Science Digital Library (cserd.nsdl.org). TeraGrid also sponsors week-long summer workshops, an annual TeraGrid summer conference, and the education program of the annual Supercomputing conference (this year in Austin, Nov 15-21; see sc08.supercomputing.org).

(91-minute audio; presentation materials for Lathrop, Doyle, Fratkin)


Panel: Integration of CI and Traditional Research Methodologies

George Askew, Associate Director for Agriculture and Natural Resources; Director of the Belle W. Baruch Institute of Coastal Ecology and Forest Science
Jill Gemmill, Executive Director, Cyberinfrastructure Technology Integration (CITI), CCIT; Research Assistant Professor, School of Computing; CU Genomics Institute (co-chair)

Elizabeth Dennis Baldwin, Assistant Professor of Parks, Recreation and Tourism Management
Neil Calkin, Associate Professor of Mathematical Sciences
Gene Eidson, Professor of Biological Sciences; Director, restoration ecology focus area and Center for Applied Ecology in the Restoration Institute; Adjunct Professor, Odum School of Ecology, University of Georgia
Patrick Fortney, Assistant Professor of Civil Engineering

Professor Baldwin described the Open Parks Grid at Clemson University, a cyberinfrastructure concept to integrate parks and people who use, manage and study them. She observed that gaps exist between researchers and practitioners, funding entities and protected areas, local and national parks, cultural and natural resources, and parks and citizens, especially children. Building on forty years of park management and education experience at the university, this interdisciplinary effort addresses parks’ geographic dispersion and management isolation, resource degradation, lack of management training, and inefficient resource utilization.

Professor Eidson reported on the Clemson Public Services Activities (PSA) Remote Data Acquisition Network Planning Committee, which focuses on watersheds at risk from population growth, economic development, and coastal urbanization. He asserted that water policy will shape the economic future of South Carolina, and argued for a statewide approach to water resources management based on real-time, sensor-based, remote data acquisition and analysis – “digital watersheds” and “intelligent rivers.” Through this recently funded effort, three projects are to be integrated – Bannockburn coastal research, Savannah River, and Clemson Forest.

Professor Fortney discussed a vision for structural engineering research and practice based on integration of telepresence and hybrid testing (i.e., combining computer physical testing, such as in wind tunnels, with computer simulation). If successful, he envisions payoffs for experimental and analytical researchers, practicing engineers, educators, and the public.

Professor Calkin, a mathematician with extensive experience using networked multimedia collaboration tools for instruction, discussed the growing impact of CI on the field. He cited the importance of electronic journals and electronic access to traditional journals, for example via JSTOR (www.jstor.org). He observed also that the field of mathematics is increasingly becoming an experimental science, and noted with approval the prospects for integration of mathematics research and education, including engagement with undergraduates and outreach to K12 students.

In the question and answer session, panelists were asked how CI could respond to their biggest research challenges. Fortney cited the need for IT expertise, Baldwin noted resistance to data sharing among some potential collaborators, and Eidson called for more and better visualization to make complex topics accessible to multiple audiences. Eidson observed also that multi-site, multi-collaborator projects require strong project management assistance, while Calkin emphasized the importance of interdisciplinary connections. Gemmill cited the HubZero effort at Purdue as an example of the importance of building reusable CI components.


(54-minute video; PowerPoint for Baldwin, Eidson, Fortney)


Report Out from Breakout Discussions & Instructions for Applying for CI Seed Grants

James R. “Jim” Bottum, Vice Provost and Chief Information Officer, Clemson Computing and Information Technology (CCIT), chair

Jeff Appling, Associate Dean for Curriculum, Office of Undergraduate Studies
George Askew, Associate Director for Agriculture and Natural Resources; Director of the Belle W. Baruch Institute of Coastal Ecology and Forest Science
Elizabeth Dennis Baldwin, Assistant Professor of Parks, Recreation and Tourism Management
Jill Gemmill, Executive Director, Cyberinfrastructure Technology Integration (CITI), CCIT; Research Assistant Professor, School of Computing; CU Genomics Institute
Barbara Weaver, Manager of Instructional Services, CCIT

Following the last panel discussion, attendees were invited to participate in one of three facilitated discussion sessions on application of cyberinfrastructure: CI in teaching and learning, CI in university research, and CI in university outreach. The closing session was devoted to summary reports from these discussions, followed by instructions to Clemson attendees for applying for CI seed grants.

CI in Teaching and Learning. Jeff Appling and Barbara Weaver reported that, overall, Clemson instructors feel well served by the facilities, services, and personnel devoted to teaching and learning support. Some of this satisfaction was attributed to the cross-campus synergy generated by the Teaching with Technology Community, a group of faculty and instructional support personnel who have been meeting weekly for nearly six years to explore new pedagogies, new technologies, and concomitant need for support services.

The discussion did identify issues and concerns, however. As more faculty members rely upon digital resources and networked delivery systems for both teaching and students’ class preparation and assignments, questions arise about CI capacity planning and reliability, together with data security and preservation. Another concern is that many faculty members are not aware of the options, tools, best practices, and support resources available; compounding the communication challenge, these options, tools, practices, and resources are evolving rapidly. Also noted are the CI challenge posed by Clemson’s distributed campuses and international programs and the opportunities for innovative instructional programs afforded by Clemson’s attractive campus, adjacent arboretum, and mild climate.

CI in Research. Elizabeth Baldwin and Jill Gemmill summarized a group discussion that focused on the promise of, and barriers to CI-supported research collaboration. Clemson enjoys a culture of collaboration, yet to achieve the full benefits of cyberinfrastructure, a set of structural factors must be addressed. One concern expressed is that, in promotion and tenure deliberations, sole authorship is valued more than joint authorship, notwithstanding the need for interdisciplinary approaches to society’s most pressing problems. Additionally, it is not clear if or how credit or acknowledgement can be earned for providing assistance or for being a valued team member.

The group offered several recommendations to address these concerns. Senior faculty were encouraged to take the lead on forming collaborations, since they know more colleagues in other units – and other universities -- and are more knowledgeable about how to work with, and around, bureaucratic processes; they, in turn, could facilitate junior colleagues’ joining the teams. Also suggested was a protocol by which successful and ongoing teams do campus “road shows” about their collaborations, perhaps as a de facto requirement for having received support from others, including from the Libraries and CCIT. Finally the group recommended a new University award for team excellence, to complement existing awards for teaching, research, and service.

Echoing the teaching and learning discussion, the research group cited a need for more information about available services and expressed concern about available capacity should numerous additional researchers respond to the newly discovered opportunities. The group called for a comprehensive directory of CCIT services as well as pointers to external services and sources of support. Also acknowledged was the need for metadata expertise and support from the Libraries. Jim Bottum commented that CCIT is exploring application of the successful “condo model” for programming support.

CI in Outreach. George Askew summarized a wide-ranging discussion of the prospects of enhancing outreach via cyberinfrastructure. The outreach group feedback on new applications, support services, and, again, the requirement for reliability. New applications of CI included:

• Real-time data acquisition and communication to and from extension agents, especially in critical incident monitoring and response
• Monitoring interns stationed remotely, around the state, the nation, and the world
• Connecting all on- and off-campus programs

The outreach group called for new or upgraded CI support services, such as:

• “How to” guides on collaboration with other universities, agencies, and client groups
• Better support for audio and video conferencing facilities, including online calendaring
• More and better desktop/laptop collaboration tools to complement the dedicated facilities
• Specialized training of help desk personnel to understand the needs of remote users
• More online resources, such as a CI blog to provide assistance on specific topics
• Exploration of ways to reach new audiences, notably those unaware of or perhaps uncomfortable with technology

The outreach group also emphasized the importance of CI robustness and redundancy. As more people and programs – especially those at a distance – depend on Clemson for mission-critical communication, information resources, and support services, the University faces a concomitant risk of program disruption and disillusionment should the cyberinfrastructure – technology and/or expert personnel – fail to meet expectations.

CI Seed Grants Program Announcement. George Askew announced that the University’s three mission vice presidents – Provost Helms, VP for Public Service and Agriculture Kelly, and VP for Research and Economic Development Przirembel -- had contributed funds for a competitive CI Seed Grants Program. With a deadline of June 6, the program sought to promote integration of cyberinfrastructure into research, teaching and outreach, and targeted multi-user workshops, symposia, and preliminary research projects. Eligible applicants included all Clemson full-time tenured or tenure-track faculty and research track faculty who participated in the CI Days Workshop. Proposed projects must involve multi-disciplinary research, teaching or outreach, with preference given to applicants who propose innovative integration of cyberinfrastructure into cutting edge activities that target Clemson’emphasis areas and assist the university’s goal to achieve Top 20 status. Preference will also be given to proposals that involve Creative Inquiry.

Closing Remarks. CI Days Program Committee Chair Barbara Weaver expressed thanks to the committee, presenters, and attendees. CIO Jim Bottum endorsed prior speakers’ observations that Clemson’s energy, enthusiasm, and collaborative spirit are unique and that the University should be proud of its rapid progress in deploying and harnessing cyberinfrastructure. The challenge remains of setting priorities, avoiding the syndrome of “never seeing an idea we don’t like,” and determining together how best to harness the energy, collaborative spirit, and CI momentum to take the University where it wants to go – into the top 20 of public research universities.

(40 minute video)

¹ See Revolutionizing Science and Engineering Through Cyberinfrastructure: Report of the National Science Foundation Blue-Ribbon Advisory Panel on Cyberinfrastructure ("The Atkins Report"), January 2003, www.nsf.gov/od/oci/reports/atkins.pdf; and NSF's Cyberinfrastructure Vision for 21st Century Discovery, July 2006, www.nsf.gov/od/oci/ci-v7.pdf

² See, for example, Cyberinfrastructure for Education and Learning for the Future: A Vision and Research Agenda, 2005, www.cra.org/reports/cyberinfrastructure.pdf

³ The Large Hadron Collider (LHC) is a particle accelerator complex recently buit near Geneva that will support collaboration by over two thousand physicists from thirty-four countries; see http://lhc.web.cern.ch/lhc/. The massive amounts of data generated by the LHC will be analyzed collaboratively using the LHC Computing Grid; see http://lcg.web.cern.ch/LCG/.

⁴ See Our Cultural Commonwealth: The report of the American Council of Learned Societies Commission on Cyberinfrastructure for the Humanities and Social Sciences, 2006, http://www.acls.org/uploadedFiles/Publications/Programs/Our_Cultural_Commonwealth.pdf

⁵ See, for example, W. Wulf, "The Collaboratory Opportunity," Science 261, August 13, 1993; and National Collaboraties: V. Cerf, et al., Applying Information Technology for Scientific Research, Computer Science and Telecommunications Board, National Research Council, National Academy Press, 1993, http://books.nap.edu/openbook.php?isbn=0309048486; and R. Kouzes, et al., "Collaboratories: Doing Science on the Internet," IEEE Computer 29, August 1996, http://portal.acm.org/citation.cfm?id=620518

⁶ Information about Clemson's Condor pool and how to use can be found at www.clemson.edu/ccit/citi/info_for_users/documentation/condor/index.html

⁷ Mark David Miliron, "Insight Initiatives," Educause Review 42, March/April 2007, http://net.educause.edu/ir/library/pdf/ERM0727.pdf

Appendix 1

Appendix 2

Appendix 3

Appendix 4

Appendix 5