It was November of 2017 when we last visited the topic of the National Strategic Computing Initiative (NSCI). As you will recall, the NSCI was started with an Executive Order (E.O. No. 13702), that was issued by President Obama in July of 2015 and was followed by a Strategic Plan that was released in July of 2016. The question for November of 2017 was, how well the E.O. would survive the transition to the Trump administration, especially since during the campaign, the new president had promised to cancel all of the Executive Orders from the previous administration. The good news about the update is that it looks like the NSCI will not only survive but will thrive.
On Friday, November 15, 2019, the White House Office of Science and Technology Policy (OSTP) and the National Science and Technology Council Subcommittee on Networking and Information Technology Research and Development (NITRD) issued the update (National Strategic Computing Update: Pioneering the Future of Computing). The update is the result of a multi-agency Fast Track committee that was convened in June of 2019. That group used, among other inputs, responses to an industry-wide Request for Information (RFI) and a Future Computing Community of Interest Meeting. The update made refinements to the NSCI that reflect changes in the computing landscape from 2015 to 2019.
From an outside federal government observer’s perspective, the update is fascinating because it reflects a level of agility not usually seen with large initiatives. Normally, once a government initiative is launched, it will “stay the course” for many years, perhaps decades, regardless of changes to the landscape. The NSCI update is a wonderful example how of initiatives should be adapted to deal with the realities of evolving technologies and programs. Some of the interesting “adjustments” include:
Exascale – In the 2016 NSCI Strategic Plan, the word “exascale” appeared over 40 times. One of the five objectives in the plan specifically talked about the need to accelerate the development and deployment of “capable” exascale computing systems. In the 2019 update, the word “exascale” is used only once, and that is in the context of the range of available computing power. The update clearly reflects the reality that OSTP and its committee consider U.S. exascale computing as “being in the bag.” This seems to be an acknowledgement of the three recent Department of Energy (DOE) contracts that will commission three different architecture exascale systems at three different national laboratories in the early 2020s.
Artificial Intelligence (AI) and Machine Learning (ML) – The 2015 E.O. and the 2016 Strategic Plan did not directly discuss AL or ML. Rather, they use the term Data Analytic Computing. Both documents indicated that there seemed to be a potential role for Data Analytic Computing that would be used with traditional HPC enabled modeling and simulation. The initial NSCI documents advocated that integration of the technologies be explored. In the 2019 update, the question of “whether” AI and ML would have a role seemed to be settled. The update moves on to advocate research into questions about “how” to deal with the added complexities of making the integration actually happen.
Beyond Moore’s Law Computing – The 2015 NSCI E.O. and the 2016 Strategic Plan spent a considerable amount of space discussing the expected challenges of approaching the limits of CMOS semiconductor technologies and how to continue to develop more powerful computers. The original NSCI documents called for establishing a “viable path forward” for future HPC systems. The 2019 update recognizes that several paths are starting to emerge (including analog, quantum, and neuromorphic computing) and advocates embracing a diversity of hardware and software approaches for the future of computing.
Cybersecurity – In the first NSCI documents, the word “cybersecurity” only appeared a few times. However, in the 2019 update, cybersecurity takes on a much more prominent role. The update provides a lengthy description of the challenges and opportunities for cybersecurity in future computer ecosystems and makes a number of recommendations.
Strategic Computing Ecosystem – The 2015 E.O. first introduced the idea that an NSCI ecosystem must exist and be nurtured to enable the effective use of strategic computing. This was reinforced in the Strategic Plan that used several pages to describe the ecosystem and to provide recommendations for of how it could be maintained and strengthened. The 2019 update discusses the ecosystem in terms of a “Strategic Foundation for Computing.” The update reinforces the importance of the 2015 and 2016 elements of the ecosystem and adds several new items. These include developing quality data used for training AI systems and the need to turn new computing capabilities into practical and usable forms for a skilled workforce.
Partnerships – The entire collection of NSCI documents place an important emphasis on partnerships. The E.O., Strategic Plan, and Update make it clear that the success of the NSCI hinges on the ability of government agencies, industry, and academia to work in a coordinated fashion. This is needed to overcome the very complex hardware, software, applications, and user environment challenges to produce future computing capabilities to address problems of national significance. The basic organization, agency roles, and responsibilities of the NSCI remain largely the same, but with the addition of a new subcommittee to coordinate cross-agency work for the future of computing at the classified level.
Overall, the 2019 NSCI update is a welcomed sign that the Trump administration understands and has adopted the idea that U.S. leadership in high-end computing provides an important strategic advantage. The question is – does that matter? The answer is a bit of “yes and no.” The “no” part is that OSTP and NITRD are not funded agencies and in reality can do very little to implement the recommendations contained in the NSCI. Their role is leadership, coordination and advocacy. The “yes” part is that funded offices and programs in agencies (e.g. the Departments of Energy, Defense, and Commerce) can point to the initiative when they make their funding requests. That is not necessarily a guarantee that they will get the request for NSCI activities, or that they will receive the appropriation from Congress, but it certainly can help. In the end, you will need to “follow the money” to understand the impact of the NSCI and its latest update. In terms of government initiatives, the NSCI is still very young, but its future looks promising.
About the Author
Alex Larzelere is a senior fellow at the U.S. Council on Competitiveness, the president of Larzelere & Associates Consulting and HPCwire’s policy editor. He is currently a technologist, speaker and author on a number of disruptive technologies that include: advanced modeling and simulation; high performance computing; artificial intelligence; the Internet of Things; and additive manufacturing. Alex’s career has included time in federal service (working closely with DOE national labs), private industry, and as founder of a small business. Throughout that time, he led programs that implemented the use of cutting edge advanced computing technologies to enable high resolution, multi-physics simulations of complex physical systems. Alex is the author of “Delivering Insight: The History of the Accelerated Strategic Computing Initiative (ASCI).”