Although initial FY17 budget proposal numbers for the first full year of the NSCI are available, they do not paint a complete picture of where the program currently stands, and based on the broad range of technical – and non-technical – challenges outlined in the original plan, there may be additional groundwork needed to ensure this program has the best chance of reaching its full potential.
Drawing on related past initiatives seen here and abroad, IDC offers the following suggestions that could help boost prospects for early and continuing NSCI success that include appointing an NSCI wizard as a government/industry focal point, amassing a series of select blue ribbon panels of US HPC industry experts to help inform NSCI policy actions, gathering HPC success stories to illustrate and justify NSCI efforts going forward, and perhaps most important, sponsoring a series of workshops that facilitate greater government industry HPC collaboration.
With the recent release of the various US Government FY2017 agency budget proposals, HPC industry pundits have been looking closely at how the National Strategic Computing Initiative fared. The NSCI was established by executive order last year and it is designed to maximize the benefits of HPC research, development and deployment across the U.S. public and private sectors. Viewed purely from a budget perspective, the early results look promising: the new federal requests call for an NSCI investments of $285 million on the DOE side, and another $33 million from the NSF.
That said, there are a number of other agencies involved in the program – eight to be exact – that have not weighed in with similar budgetary details, and as a result, it is not possible to determine where the full NSCI funding call across the entire US Government currently stands. For example, there is not much mention of NSCI efforts within various deployment agencies such as the Department of Homeland Security or the FBI.
Because these agencies are relatively new to the HPC world compared with those such as the NASA and NOAA how these organizations take their first steps into this effort will bear close observation. In addition, for cases where NSCI line items are evident, special attention must be paid to ensure that new NSCI work gets new NSCI money. Instances where long-standing efforts are simply relabeled NSCI or are cannibalized to support NSCI need to be ferreted out and identified as such to ensure that the NSCI funding is accurately represented.
Adding to the uncertainty surrounding the current true state of NSCI support is that this is a merely a budget proposal for a fiscal year that starts essentially at the end of the current administration. Many view much of this year’s budget requests as either wish lists or at best first drafts of plans that will no doubt undergo some significant alterations down the road. And it is not yet clear the extent that the NSCI is viewed either favorably or unfavorably on the Hill. The best conclusion anyone can really draw now is that the various agencies have moved to secure budget for NSCI-based work going forward, but only time will tell if the commitment of resources will be forthcoming.
But that is really only part of the problem. Watching this anxiously from the sidelines are a number of U.S. commercial HPC-related interests, both within the supplier community and the base of HPC users, current and potential, who are looking to participate in the NSCI effort. Discussions with some in the commercial sector indicate that there is not much news to go on and that they are awaiting some outreach plan from the U.S. Government. These commercial entities indicate that now would be an opportune time for the various federal agencies to formulate and soon execute plans to engage with industry counterparts. Indeed, some may argue that this aspect of the program could ultimately be the make or break determinant for the entire NSCI effort.
To understand this, it is important to remember some of the critical goals of the NSCI as laid out in its original vision:
- Foster a robust commercial HPC sector.
- Keep the U.S. as the leading supplier nation.
- Broadly deploy HPC capability in the US private.
- Build an HPC workforce.
- Train wide range of non-HPC scientists and engineers use or improve their use of HPC technology.
It is clear that these goals cannot be realized by the various government agencies working alone or even collaboratively. Instead, they likely will require an enhanced outreach program from the various US government agencies to begin the process of a building a stronger US government/industry collaborative infrastructure. Doing this won’t be easy: it is not an area where either U.S. HPC mission agencies or HPC-related corporate entities have significant experience, and in addition, there will be many tough legal, technical, and political issues to iron out before the process can reach its full potential.
Drawing on related past initiatives seen here and abroad, IDC offers the following suggestions that could help boost prospects for early and continuing NSCI success.
For its part, the U.S. Governments should consider appointing an HPC wizard who can serve as the government face of this effort. As it now stands, OSTP is running the NSCI from a strategic perspective, but the effort could be well served by identifying a person closer to both the government and commercial HPC development world who can easily move between the two, making connections and smoothing the process, while acting as a lightning rod for questions and answers.
In addition, the government could look to draw on the collective talents of a panel of industry experts who can help consult with OSTP and related agencies on how best to move forward from here. The initial executive order called for the creation of an executive council that could interact with the private sector, and it may be time that effort began in earnest.
Meanwhile, OSTP, or one of the developing agencies, should consider running a series of government industry workshops that can serve a number of purposes: get all the right people in the same room talking and sharing ideas, help to establish linkages between the various missions agencies and potential industry partners, and finally solicit ideas from the commercial sector as to how best to build a collaborative infrastructure that highlights some of the challenges with – and perhaps suggest changes to – current regulations or laws dealing with government/industry partnerships.
Finally, these workshops could be leveraged to gather and share HPC success stories that can be used to clearly illustrate and justify continued funding for the NSCI with an eye towards demonstrating progress in both scientific development and economic competitiveness
IDC has stated in the past that simple funding efforts that move to address the single mission requirements of various HPC mission efforts will not be enough to help drive the U.S. HPC sector as a global technology leaders nor will it provide sufficient uptake for HPC technology into overall U.S. industrial sector. To achieve these challenging but attainable goals, the NSCI likely should soon initiate the process whereby the US Government and commercial sector start to lay a foundation for a strong, enduring collaborative infrastructure.
The most direct way forward to achieve this is not clear, but bringing the collective intellect of both government and commercial HPC leadership to bear on the problem offers the best chance for a solution. Ultimately, IDC believes that establishing the opportunity for meaningful dialogue in the coming months could go a long way toward breaking down barriers within the U.S. Government HPC community as well as fostering a much more collaborative environment between the U.S. HPC government and the commercial sector. This is a not a nice-to-have element of the NSI; it’s a must-have element in realizing some of the key visions of the program that could help cement the U.S. as a leading HPC developer and user nation for the rest of this decade and beyond.
Links to more discussion of NSCI’s progress in recent HPCwire articles:
Budget Request Reveals New Elements of US Exascale Program, http://www.hpcwire.com/2016/02/12/obama-budget-reveals-new-elements-exascale-program/
Final Obama Budget Lauds Innovation, Unlocks NSCI Funding, http://www.hpcwire.com/2016/02/10/final-obama-budget-fy17/