2 Strengthening creative research in the university community

Indiana University has long been a national leader in high performance computing for research. IU faculty and student researchers in areas from astronomy to the arts and humanities have no-cost access to powerful supercomputers and massive data storage critical to their research.

RT staff strive to enable researchers to reach discovery faster with low barriers to their research. From education and outreach to tools like Research Desktop (RED) and the first NSF-funded cloud for science and engineering research, Jetstream, we make it easier to access vast amounts of storage, research software, and fast calculations on IU’s supercomputers.

By the numbers

Science gateways on Jetstream served more than 100,000 researchers and students as of June 1, 2022.

Jetstream2 began early operations in February 2022 with 8 petaFLOPS of virtual supercomputing power and 17 petabytes of storage.

The Scholarly Data Archive can hold enough data to fill about 225 tons of DVDs.

The Slate high performance file systems received 100 petabytes of data last year from Big Red 200—a constant transfer rate of 3 gigabytes per second for 365 days.

Researchers have run  508,811 jobs on Big Red 200 since it was made available in spring 2022, using 197,336,920 core hours. 

In 90 seconds, Big Red 200 could . . .  

• Do more calculations than there are stars in the Milky Way (10^11–10^12 stars).  

• Complete 2.7 million years’ worth of credit card transactions. 

• Figure about the same number of calculations as there are unique NCAA tournament brackets (2^63 or 9.2 quintillion). 

• Perform the number of operations that, if each operation were a kernel of corn, would take the state of Indiana 12,000 years to produce.

Highlights

Developing virtual collections
IU3D and the Advanced Visualization Lab have partnered with a variety of IU collection holders to design and implement innovative online and in-person displays for presenting and interacting with digital versions of their physical assets. Partners include Collections @ IU, the new Museum for Archeology and Anthropology, the Eskenazi Art Museum, the Lilly Library, the Sage Costume Collection, the IU Bicentennial, and the Galleries at Herron School of Art + Design.

Cataloging the stars
Using the Data Capacitor II high-speed shared storage system and the Carbonate supercomputer, IU astronomy researcher Caty Pilachowski and her team were able to analyze and catalog 70,000 stars in the Milky Way’s galactic bulge. The resulting catalog is the largest ever produced and will be made available to astronomers everywhere.

Streamlining genomics data management
The IU School of Medicine Center for Medical Genomics and the RT Scalable Compute Archive (SCA) team collaborated to lower the barriers of complicated information technology systems and workflows.

“The solutions of the SCA allow us to enhance our data management practices by streamlining the data processing, data dissemination, and data backup with just a few clicks. In addition, the users can also easily share their data with partners within the system, which enables collaboration.”

—Yunlong Liu, director, Center for Medical Genomics, Indiana University School of Medicine

Helping IU researchers leverage cloud computing
Research Data Services partners with multiple IU research groups leveraging commercial and on-premises cloud computing resources. These groups include researchers from the Luddy School of Informatics, Computing, and Engineering; the Kelley School of Business; the IU Network Science Institute; SPEA; and the IU School of Medicine. These partnerships include implementations on Jetstream as well as integrations on all three major commercial cloud platforms.

“The Alcoholic Hepatitis Network consortium, with its 10 institutional partners, is counting on us to make the ARDaC cloud platform work—and RDS made it happen.”

—Jing Su, assistant professor of biostatistics & health data science, IU School of Medicine

Supporting data collection and management
IU’s Research Electronic Data Capture (REDCap) is a secure, web-based platform that enables researchers to build and manage online surveys and databases with no programming experience. Researchers across the state have used REDCap to:

• Collect and store data about Indiana homes affected by lead.
• Design and build behind-the-scenes research and operations data capture for the IU School of Medicine/AstraZeneca COVID-19 vaccine trial.
• Evaluate the impact of COVID-19 on the mental health, addiction, and help-seeking behaviors of American Indian and Alaska Native young adults.

“Social network data are complicated to collect, in general, and in the survey format. We use REDCap because it has the capability of dealing with the challenges of both network generators and name interpreters as well as the novel mental health items on suicide risk and resilience.”

—Bernice Pescosolido, distinguished professor of sociology, Indiana University

Understanding the risk of COVID-19 on cancer patients
The biomedical informatics field is under increasing pressure to analyze large-scale health data quickly to inform policy and develop interventions. Researchers at the Regenstrief Institute used Slate-Project to examine data to help them better understand the effect of COVID-19 on immunosuppressed cancer patients in Indiana.

“The Slate-Project environment is perfect for analyzing the larger datasets necessary to infer populations trends and outcomes. Storage helps with large datasets and performance helps us do our work quickly and efficiently.”

—Brian Dixon, director of public health informatics, Regenstrief Institute

Researching one billion years of cosmic history
Francesco Calura, an IU collaborator from the Astrophysics and Space Science Observatory of Bologna (OAS) used Big Red 200 to model the formation of globular clusters. Big Red 200 has the capability to research approximately a billion years of cosmic history using a set of simulations with unique features.

“[A]fter a few weeks of intensive use on Big Red 200, my simulations have moved forward signi cantly at a pace that is unprecedented for what I have seen on IU machines (and anywhere else).” 

—Francesco Calura, researcher, INAF-OAS, Astrophysics and Space Science Observatory of Bologna: Bologna, Emilia-Romagna, IT

Enabling new black hole research
Astronomers on the Event Horizon Telescope team released the first image of Sagittarius A*, a black hole at the center of the Milky Way. The workflows used to capture the image were protoyped on Jetstream.

“Jetstream kick-started the EHT’s cloud computing effort. The EHT now can survey through hundreds of thousands of imaging parameters because of our initial work with Jetstream.” 

—Chi-kwan Chan, leader of the EHT Computations and Software Working Group

Transferring power wirelessly
As governments and industry leaders seek alternatives to fossil fuels powering manufacturing and large-scale electric grids, IUPUI Professor and Director of the Richard G. Lugar Center for Renewable Energy Peter Schubert looks to space as a reliable source of solar power. He used Big Red II, Big Red 3, and Carbonate supercomputers to model sending beams of solar power to Indianapolis via satellite.

“The ionized atoms, the electrical fields are very sensitive and the elements want to shoot out in all directions, so if you’re dealing with a wisp of vapor, you can model them as alone in space. But once you start packing them into a dense beam, the interactions multiply—so it very quickly snowballs into a calculation that’s too much for a laptop. Coordinating the simulations, software, and systems is a complex process, and [Research Technologies] did really great work to help us solve really hard problems.”

—Peter Schubert, professor of electrical and computer engineering and director of the Richard G. Lugar Center for Renewable Energy at IUPUI

Managing massive amounts of data
Through our Research Data Services (RDS) group, we are developing and deploying broadly applicable workflows for data and metadata management that are helping individual labs, science departments, and research-intensive schools to manage their research data in consistent, secure, and sustainable ways. These services leverage the excellent resources of RT’s Research Storage and High-Performance File Systems teams.

“RDS helps us manage, store, and process large amounts of imaging data produced in our labs.”

—Hui-Chen Lu, Gill Chair of Neuroscience, Department of Psychological and Brain Sciences

Improving child well-being
Dashboards help community leaders identify adverse trends and address them quickly to improve child well-being. RT’s Crisis Technologies Innovation Lab and IU’s Biostatistics Consulting Center collaborated with The SOURCE, Elkhart County’s system of care, to develop a dashboard that was prototyped on Jetstream. The dashboard provides real-time information about trends and well-being, risk, and protective factors affecting children in the community.

“You have to convince people to share data and then continually reinforce that you’re not going to use that data to point fingers. There’s this continual balance of trust and reinforcing those messages all the time with people.”

—Rebecca Shetler Fast, director, Elkhart County’s system of care, The SOURCE

Enabling new statistical tools
The RT Research Applications and Deep Learning (RADL) team worked with the Bloomington Assessment and Research (BAR) office to create a proof of concept tool and associated test environment to allow BAR to expand opportunities to deploy analytics developed in the R statistical language.

“This application is a prime example of how strategic use of data, software, and analytics can help improve student success here at IU.”

—Sabrina Andrews, associate vice president for institutional analytics

Ensuring ethical animal research
IU Bloomington School of Public Health researchers conducted National Institute of Aging–funded research on the Karst supercomputer to find which test yields optimal statistical results with proper error rates using the least number of animal subjects. The simulations were based on data from real mice to better reflect how real-world data would behave. 

“That’s one of the huge benefits of using supercomputers. They enable us to do the calculations more than once in a reasonable amount of time to make sure we’re getting the right answers.”

—Andrew Brown, assistant professor, Department of Applied Health Science, School of Public Health

Innovating information retrieval
Titus Schleyer (Regenstrief Institute) and Xia Ning (The Ohio State University) are using Carbonate and Slate-Project to store and process data for two projects: an innovative information retrieval method for electronic health records using collaborative filtering and a project to characterize COVID-19 patients through a community health information exchange and electronic health record databases. Both projects require access to a large amount of data that contain protected health information.

“The good thing about Carbonate and Slate-Project is that they provide mechanisms to protect sensitive information. They also have ways to monitor for a possible data breach. This is very helpful for us, so we use Carbonate and Slate-Project for both of our research projects.”

—Xia Ning, associate professor for biomedical informatics and computer science and engineering at The Ohio State University

Connecting researchers around the globe
IU Libraries, the IU Network Science Institute (IUNI), and the Big Ten Academic Alliance partnered to create the Collaborative Archive & Data Research Environment (CADRE), an environment where researchers can access big datasets, reproduce shared results, and replicate shared queries. Originally designed and built in Jetstream, CADRE has two components. One is an on-premises system, which works on local Carbonate nodes, but the majority of CADRE is in a cloud-based gateway.

“By encouraging these types of collaborative relationships, we’re not duplicating work and we’re creating a more open environment and encouraging more data to be open so that more researchers can work on it, collaborate, and share work which in turn should increase reproducibility.”

—Jaci Wilkinson, executive director of CADRE