The Texas A&M University System and IBM have created one of the world’s largest computational sciences infrastructure. Built on a mix of high performance computing technologies that include Blue Gene/Q, Power 7 and System x servers, scientists and engineers across the system’s 11 universities and seven state agencies will work on projects dedicated to advancing agriculture, geosciences and engineering. Early tests pitted the Blue Gene/Q, installed at TAMUS’s flagship campus in College Station, on a material sciences problem that previously took weeks. The Blue Gene/Q’s 418 Teraflops solved it in less than an hour.
Jon Mogford, the Vice Chancellor of Research for TAMUS, will play a key role in the coordination and success of these diverse projects, teams and technologies. He met with IBM Smarter Planet to discuss how the university system is putting this massive new infrastructure to work.
A Smarter Planet: How does the new infrastructure improve the university system’s ability to conduct geoscience, engineering, and other research?
Dr. Mogford: The new infrastructure greatly improves our computational capacity and the ability to assimilate and manipulate large data sets from a variety of sources. This in turn will allow refinement of models, design algorithms and testing in a broad collection of applications such as polymer modeling, genomic signal processing, ocean circulation and climate models, among many others. This, of course, is great — however, we are more excited about the enterprise-wide research partnership agreement, which really drove the entire discussion between A&M and IBM.
This enterprise-wide relationship emerged from a series of capability meetings over the last year that involved research leaders from both organizations. The discussions spanned the fields of Engineering, Geosciences, Agriculture, Veterinary and Human Medicine. We challenged ourselves to consider the grand challenges in all of these fields and how the partnership could bring about tangible outcomes. We are highly complementary groups from the R&D perspective, and that has come through in all our planning meetings.
A Smarter Planet: The system’s universities and state agencies are spread across Texas. How will they tap into these new computing assets?
Dr. Mogford: The regional universities will tap into these new computing assets via our TTVN/Learn network. All of the regional universities are connected to a dual 10GB backbone that the research data center will be resident on. This backbone is the same network that allows the regional universities to access system assets for payroll and financial systems.
A Smarter Planet: How does the big data of the mentioned research areas go from the field to the computing system for analysis?
Dr. Mogford: This is a challenge due to the geographic spread, data structure and breadth of discipline and research ongoing in the A&M System. Obviously, the data will need to be transferred across the network or via transportable media from the originating source to the research data center’s storage array or servers.
The decision on how to do this will depend upon the amount of data, the originating source of the data, and complexity of data to be merged. But challenges are also opportunities. Barriers to field collection and transfer of large data sets — in the appropriate structure — will certainly benefit from the attention the Research Partnership.
A Smarter Planet: What are some of the first things you want to accomplish as your engineers and scientists begin using the technology?
Dr. Mogford: The research partnership meetings have already led to detailed discussion in a wide variety of fields. Tracking of animal disease, modeling and design of nanoscale carbonized materials, smart grid, genomics signal processing, and emergency informatics are just a few examples.
From our College of Geosciences, priorities will be given to enhancing the SmartGulf effort, which is focused on real and near real-time observations of the Gulf of Mexico, and the modeling of these data for policy makers. Additional efforts will focus on the Texas Water Observatory in an effort to develop a real and near real-time smart network to analyze/monitor hydrologic system for the State of Texas. Efforts will also focus on use of large datasets for petroleum exploration and basin analysis.
Finally, the newly required infrastructure will allow the development of a computational geosciences curriculum. Industry has a growing need for graduates who have experience in high-performance computing. Having these facilities will help the university recruit and retain the best students.
A Smarter Planet: What should potential students, from undergraduates to PhD candidates, who are interested in these areas of study consider making part of their academic curriculum?
Dr. Mogford: Regardless of how one defines “Big Data,” there is no doubt that our ability to realize, detect, collect, transfer and store data is growing at a rapid pace. It would certainly be difficult for students to not be aware of this, but they should definitely understand the opportunities provided by this growth – and have an even better appreciation of the limitations and challenges at the analytical and decision stage.