The National Strategic Computing Initiative - A Not-So-New Program

Posted by Alexander Rossino on August 18, 2015

Small Businesses in Big Data

On July 29, 2015, the White House published an executive order outlining its priorities for a National Strategic Computing Initiative (NSCI). The NSCI, which calls for investments by several executive branch agencies to accelerate research and development in supercomputing, was received with great public fanfare despite the fact that the NSCI is not new. The policy foundation for the initiative has been under development since at least 2012 and several major procurement actions have already been completed for it. This leads to the following question – if the NSCI was already in play why was an executive order mandating it necessary? The answer is the budget deadlock between Congress and the White House. Back in February 2015, when the president’s budget request for fiscal year 2016 was issued, the Analytical Perspectives volume of the request included $242 million for the National Strategic Computing Initiative as part of the Networking and Information Technology Research and Development program. Since that time, the budget has gone nowhere in Congress. Furthermore, the prospects of the budget passing and being approved by the president continue to look poor, so the White House took the initiative (pun intended) and issued an executive order directing the relevant agencies to invest in supercomputing-related R&D.

Now that the initiative has been announced, the question becomes where to find business opportunities related to it. Here the way ahead is trickier to navigate. According to the NSCI EO, the agencies involved are divided into three groups: 1) Lead Agencies that are “charged with developing and delivering the next generation of integrated HPC capability,” 2) Foundational R&D Agencies that are “charged with fundamental scientific discovery work and associated advances in engineering necessary to support NSCI objectives,” and 3) Deployment Agencies that “will develop mission-based HPC requirements to influence the early stages of the design of new HPC systems.”

The Lead Agencies are the Department of Energy (working in tandem with the National Nuclear Security Administration), the Department of Defense, and the National Science Foundation. Leaving aside the DOE/NNSA nexus for a moment, the NSF is directed to “play a central role in scientific discovery” while the DoD will continue to “focus on data analytic computing to support its mission.” These vague formulations basically state that the NSF and DoD will continue to perform R&D in lockstep with their existing mission requirements.

As closely as I can identify them, the NSF’s scientific discovery programs are related to its “Cross Foundation Investments” in brain research, the risk to and resilience of critical infrastructure presented by extreme events like hurricanes and earthquakes, and modeling the connections between interlocked food, energy, and water systems (See table below).

The requested NSF budget for FY 2016 shows considerable growth in all three programs, with INFEWS being an entirely new program. It is impossible to determine the contractor addressability of these budgets at this point. Vendors should keep in mind, however, that the NSF pursues a high percentage of its research through grants, not contracts.

Concerning high-performance computing investment at the DoD, the picture there is considerably less rosy as the DoD’s investment in HPC has been stable to falling for the last three fiscal years. The good news is that of the $81 million the DoD requested for HPC in FY 2016, $64 million is Development, Modernization, and Enhancement (DME) funding, also known as “new money.” The extent to which this money trickles into “big data” investments related to HPC is difficult to determine. Under “High End Computing,” the NITRD budget documents state only the following concerning DoD: “advancements in modeling and simulation of challenging environments, algorithm and technology development, and big data applications.” Fortunately, I’ve covered planned DoD investments in big data R&D here and here, so for additional insight, please see those posts.

The real spending action has been at the DOE, where over the last 2 years, the agency has awarded contracts for new supercomputers and related R&D totaling $625 million. The contracts in question are for the Collaboration of Oak Ridge, Argonne, and Lawrence Livermore (CORAL) program ($325 million), the FastForward2 program ($100 million), and the Aurora Supercomputer program ($200 million). These contracts will provide the new supercomputing resources that are necessary for the NSCI to proceed.

Meanwhile, the Foundational Research and Development Agencies for the NSCI are the Intelligence Advanced Research Projects Activity (IARPA) and the National Institute of Standards and Technology (NIST). According to the EO, “IARPA will focus on future computing paradigms offering an alternative to standard semiconductor computing technologies [while] NIST will focus on measurement science to support future computing technologies.” No budget is available for IARPA, so that provided by NIST will have to suffice. Presumably, NIST’s efforts will be carried out under the NIST Central IT Support for Science program. NIST requested $6.1 million for this program in FY 2016, an increase of $130 thousand from FY 2015’s funding of $5.97 million. As you can see, we’re not talking about a lot of money here.

The most intriguing aspect of the NSCI from a vendor perspective is the role played by the five Deployment Agencies. These agencies are the National Aeronautics and Space Administration, the Federal Bureau of Investigation, the National Institutes of Health, the Department of Homeland Security, and the National Oceanic and Atmospheric Administration. NASA and NOAA already have well-developed HPC programs with significant investments (See table below) and while NASA’s HPC budget has not fared well recently, NOAA’s has done better with $11 million of the $83 million requested for FY 2016 being DME dollars.

The NIH, FBI, and DHS do not have HPC programs. The level of funding we’re talking here for the integration and use of HPC technology by these three agencies is unknown, but the fact that they are new to the use of HPC suggests that the opportunity for vendor-provided services is potentially greatest here. The NSCI EO states that these agencies “will also have the opportunity to participate in testing, supporting workforce development activities, and ensuring effective deployment within their mission contexts.”

If this language doesn’t cry out for industry support, I don’t know what does. The $64,000 question is will the vendors sought for help be those delivering the new systems to DOE or will they be others contracted separately at the respective agencies? Those seeking an answer to this question should keep their ears to the ground at the respective agencies to jump on prospective services procurements related to the initiative.

Categories