To take on digital transformation for its 27 separate institutes and centers, the National Institutes of Health is looking ahead another five years.
The agency within the Department of Health and Human Services has a nearly $40 billion budget, about 80% percent of which goes to nearly 2,500 universities and medical centers nationwide. NIH supports about 300,000 researchers, which means, according to CIO Andrea Norris, much of its activity focuses on “investigator-researcher-initiated ideas and hypothesis for discovery.”
Norris is also director of the Center for IT — one of the 27 entities of NIH — and her office receives about $1 billion annual to maintain a core computational infrastructure. She said they move about 6 petabytes of data per day through the system.
“Our backbone network is one core because the predominant data that we’re moving through there is scientific by far, but we do have set up in some cases, we architect virtual local networks for specialized scientific instrumentation,” Norris said on Federal Monthly Insights — Digital Transformation. “We also have a scientific — what’s referred to as a scientific DMZ that can be used for dedicated needs for fast traffic, for science purposes.”
Norris still has to work with the other institutes’ CIOs, and the Federal Information Technology Acquisition Reform Act can make this somewhat of a challenge. When FITARA was first established, she said, the Center for IT worked closely with key department leadership how to implement the rules in a way that reflected HHS organizations’ diverse missions, from the Centers for Disease Control and Prevention to the Office of the Secretary.
“We present our budget requests. We work with the department each year in a prospective way on what the big IT investments are,” she said on Federal Drive with Tom Temin. “We work very closely on our acquisition strategies. Certain thresholds are dealt with at the at the operating division level or local level.”
NIH got a taste of digital transformation four or five years ago when the agency modernized its network for a tsunami of “big data,” she said. For NIH’s Biowulf, the world’s largest supercomputer dedicated to biomedical research, her office brought in outside help to determine the best kind of network architecture. It wasn’t enough to simply look at what the Energy Depart or National Oceanic and Atmospheric Administration do for their own supercomputers, she said.
“So we have more than 600 different applications that our researchers are using,” Norris said. “We provide a lot of very personalized support of expert folks who can help these researchers tune their algorithms structure their data in a way that will optimize both the compute and the power of what they’re doing analytically.”
Even as NIH wraps up this round of digital transformation it’s already begun developing the strategic plan for the next five years. The agency has a charge-back model for different institutes based on their usage and storage. A buy-in model also exists for particular researchers and institutes to acquire computational nodes and resources, for priority use but which is still managed centrally.
“We have challenges with making our data what we call fair, findable, accessible, interoperable and reusable. As we funded research activities over these many years, the data is typically stored when researchers computer, perhaps at a university or medical center data center,” she said. “And so it’s in a variety of different places, and areas. And it’s hard for other researchers to know how to access that data, is it useful, useful for them in their own research activities, and then how to make that and leave that for the next researcher.”