Advanced Computing in the Age of AI | Sunday, July 21, 2024

Digital Manufacturing Report Card: Who’s Finding the Missing Middle? 

Numerous initiatives from government and industry have been created to address the "Missing Middle" — small manufacturers who have not adopted the digital manufacturing techniques of computer-aided modeling and simulation. In this watchdog report, Addison Snell of Intersect 360 Research grades the endeavors based on what they've accomplished so far and their prospects for future success.

Numerous initiatives from government and industry have been created to address the "Missing Middle" — small manufacturers who have not adopted the digital manufacturing techniques of computer-aided modeling and simulation. Intersect360 Research has studied this area extensively and has been closely involved with several of these initiatives. In this watchdog report, I give my grades on what they've accomplished so far and their prospects for future success.

binoculars concept artAn Intersect360 Research study revealed that 62 percent of manufacturing companies with over 10,000 employees were using high performance computing (HPC) systems to assist in computer-aided engineering, but that figure was only 8 percent for manufacturers with 100 employees or fewer. In total, only about one in 12 US manufacturers employs digital modeling and simulation techniques. This problem is frequently referred to as the "Missing Middle." The term "Missing Middle" is a misnomer in one sense — I'm not sure what these organizations are supposedly in the middle of — but from a digital manufacturing perspective, they are certainly missing. And finding them is only the first step.

A number of efforts have been launched to address the Missing Middle problem. These efforts come with varied incarnations and motivations. Some are driven by government, seeking to improve national competitiveness, while others are driven by profit-seeking technology vendors. Some are individual efforts; some are industry consortia. They have had varying degrees of success to date, and they certainly have different outlooks for future success.

As industry analysts, we have researched the dynamics of entry-level and mid-range adoption through multiple projects over several years. I have been personally involved in many of the efforts below as an advisor or as a conspirator. What follows is my progress scorecard for existing digital manufacturing initiatives, rating how well they address the Missing Middle technology gap, their progress to date, and their prospects for the future.

The Alliance for High Performance Digital Manufacturing

Addressing the Missing Middle: A
Success to Date: A
Future Outlook: C

The Alliance for High Performance Digital Manufacturing (AHPDM) is a loosely-organized multi-organization consortium that meets weekly to discuss issues concerning digital manufacturing adoption. Membership is open to any organization, and the AHPDM has included hardware vendors, software vendors, government liaisons, industry analysts (including Intersect360 Research), publications (including Tabor Communications), and most importantly, manufacturers. Although no single organization is "in charge," the National Center for Manufacturing Sciences (NCMS) and Intel have contributed the most to spearhead AHPDM. (NCMS hosts the AHPDM website, and Intel coordinates the weekly call.)

To date, the AHPDM has done more than any other group or initiative to synthesize action in directly addressing the Missing Middle. Acting as a group, the AHPDM worked through NCMS to influence legislation in Congress, including the addition of specific language to the COMPETES Act. Members of the AHPDM have met with leaders in the Department of Commerce and the Manufacturing Extension Program (MEP) to discuss new initiatives. The AHPDM also provided momentum for NCMS to launch its digital manufacturing Strategic Interest Group. The AHPDM had a presence in the Ohio Supercomputer Center booth at SC10, and there are plans to repeat it at SC11.

Moving forward, the AHPDM will serve better as a launching platform for initiatives than it will in carrying anything out itself. Without a clear leader, there does not seem to be a central, agreed-upon goal for the group beyond discussion and evangelism. While the AHPDM remains important and worthwhile for any organization interested in digital manufacturing, most of the real change will be implemented elsewhere.

NCMS: Digital Manufacturing Strategic Interest Group

Addressing the Missing Middle: A
Success to Date: B
Future Outlook: B

As a not-for-profit consortium for manufacturers, NCMS is uniquely positioned to take a leadership role in digital manufacturing initiatives, and over the past year NCMS has raised its visibility and involvement considerably. This year NCMS has crystallized its role through the creation of a Strategic Interest Group (SIG) for digital manufacturing.

Building on the work begun by the AHPDM, the SIG provides a platform for outreach and awareness of digital manufacturing to thousands of US manufacturers. Any NCMS member may join the SIG, with membership fees scaling based on the size of the organization.

After announcing the intent to create the SIG at its Ann Arbor, Michigan, headquarters in March, NCMS hosted its first SIG meeting in Troy, Michigan, in July, attended by manufacturers and technology providers that had become the founding SIG members. The agenda consisted of discussion focused on "obtaining the required resources to implement Predictive Innovation Centers (PICs)" [see below] and "developing the effectively communicate to the missing middle." The next SIG meeting is scheduled for October 26-27 in Boston.

One of the benefits of the NCMS digital manufacturing SIG is that, similar to AHPDM, it has the ability to attract members from the vendor community, the manufacturing community, and the academic community. What sets it apart from AHPDM is its ability to drive forward a vendor-neutral agenda with specific goals. The eventual success of the SIG will depend on NCMS's ability to reach enough manufacturers to sustain a movement and to simultaneously deliver demonstrable progress on action items like its proposed PICs.

NCMS: Predictive Innovation Centers

Addressing the Missing Middle: B
Success to Date: C
Future Outlook: B

NCMS's primary strategy for giving SME manufacturers access to digital manufacturing capabilities is the development of Predictive Innovation Centers (PICs). Each PIC would be an HPC-capable resource for manufacturers who have a desire or need to run CAE applications but without the internal resources or scale to do so. Paying on a utility-model, core-hour basis, manufacturers could run their jobs without a large internal sunk cost.

As with the SIG, NCMS is well-positioned to lead the PIC effort. Outreach and awareness will be a challenge, and NCMS appears to be equipped to handle it. Furthermore NCMS can attract technology providers from the SIG to contribute to the PICs.

However, the challenge of addressing the Missing Middle goes beyond providing access to technology. To be complete, the PIC strategy will also need to provide manufacturers assistance and experience in using digital manufacturing. Although this is part of the PIC plan, the actual strategies for doing so are not yet synthesized. The notion of creating an HPC appliance or website that completely removes the entry barrier seems farfetched at best. Academic institutions show some promise in perhaps playing a role to contribute expertise.

Another challenge of the PIC is that even if it is successful, it could find itself in competition with offerings from competitors. R Systems, Penguin Computing and others already provide remote HPC capabilities, to say nothing of pure public cloud offerings such as Amazon EC2. The success of PICs will therefore depend on not only providing lower-cost access to scalable hardware and software, but also manufacturing-specific expertise to ease the transition to digital manufacturing.

National Digital Engineering and Manufacturing Consortium

Addressing the Missing Middle: D-
Success to Date: C
Future Outlook: C

In March the U.S. Department of Commerce awarded the U.S. Council on Competitiveness a $2 million grant "to create a public-private partnership that will help small and medium sized manufacturers in the Midwest compete in the global economy through the use of 'game changing' modeling and simulation technology," according the Council's press release. This grant resulted in the Council's formation of a Midwest pilot for the National Digital Engineering and Manufacturing Consortium (NDEMC, pronounced "endemic").

The goals proposed by NDEMC are similar to those proposed by the NCMS PIC strategy. The Council's press release stated, "The goals of the project are to lower the barriers for small and medium-sized manufacturers for adopting modeling and simulation and high performance computing for purposes of increasing their design and innovation capacity, shortening time-to-market for products, manufacturing safer products, and increasing quality while lowering costs." Beyond that statement, it is difficult to find what NDEMC actually proposes to do, or for whom. And whereas NCMS has a clear leadership role in its own initiatives, it is unclear what purpose the Council is serving, beyond being the organization that deals the money.

Once the grant was awarded, the Council called a hasty "re-launch" of its HPC Advisory Council (of which Intersect360 Research is a member), which then met at Lawrence Livermore National Labs. (No subsequent meeting has been called.) During this meeting, there was no agreement as to the profile of manufacturer being targeted, nor as to any ideas for reaching them.

For the implementation of NDEMC, the Council is being guided by insights provided by large manufacturers on their steering committee, including Procter & Gamble, John Deere, Lockheed Martin, and General Electric, who put up matching funds to the NDEMC grant. Given the financial stake of the large manufacturers involved, it is no surprise they will have a significant stake in how the funds will be used. In this case, the low-hanging fruit is the enablement of the supply chain of the contributing manufacturers to reach greater HPC scale. In other words, NDEMC will do more to help the growth of HPC within a company that already has it than it will to introduce digital manufacturing to an SME that has no experience.

This is an admirable goal — from an HPC perspective it is even desirable — but it is only loosely connected to the Missing Middle concept. NDEMC will conform to the letter of its stated intent, but it will help existing HPC users more than new ones.

America COMPETES Reauthorization Act

Addressing the Missing Middle: A
Success to Date: F
Future Outlook: Incomplete

In January Congress passed the America COMPETES Reauthorization Act of 2010. The legislation in toto outlines initiatives to better improve US competitiveness in global markets. As a direct result of efforts by NCMS and AHPDM, the COMPETES Act included Section 605, "Promoting Use of High-End Computing Simulation and Modeling by Small- and Medium-Sized Manufacturers."

This section clearly outlines the "Missing Middle" advanced technology gap and directs the federal government to take action. In particular, the Secretary of Commerce was directed to conduct, within 30 days of the Act's effect, "a study of the barriers to the use of high-end computing simulation and modeling by small- and medium-sized manufacturers in the United States." The Secretary of Commerce was further directed to provide a report to Congress on that study within 270 days. Pilot programs and demonstrations are specifically authorized by the Act.

The COMPETES Act was signed in early January, and the 270-day period is elapsing now. Intersect360 Research, the AHPDM, and NCMS have repeatedly reached out to the Commerce Department, and I personally have been involved in teleconferences or face-to-face meetings with the National Institute of Standards and Technology (NIST) — the agency within Commerce where the COMPETES Act authority allegedly resides — including representatives from the Manufacturing Extension Partnership (MEP) office within NIST.

It seems no one in the federal government — in Commerce or elsewhere — has taken any ownership of the COMPETES Act, or at least of this Section. This leads us two possible scenarios. A pro-HPC program viewpoint will see this as an excellent program, with hope that there is indeed action taking place behind the curtain, and that we will still see worthwhile initiatives coming forth from COMPETES. Contrarily, a free-market, fiscally conservative observer might hope or expect that the directives and deadlines have been ignored, passed around as someone else's responsibility until forgotten, and that nothing of substance will ever come to pass.

Vendor-Specific Marketing Initiatives

Addressing the Missing Middle: C
Success to Date: D
Future Outlook: B

It was US President Calvin Coolidge who concisely quipped, "The business of American business is business." To that end, many HPC technology vendors have a business interest in unlocking the Missing Middle, essentially creating a new potential market opportunity.

These vendor-specific initiatives are varied, in grand multi-vendor partnerships at one extreme and isolated field-office efforts at the other. The grades for this section therefore represent industry averages, with some vendors doing better than others.

The weakness in many vendors' plans is that they fall into the trap of focusing primarily on low-cost access to hardware. Solving the problem of changing a manufacturer's workflow to create digital models and to incorporate simulation is a harder and no less significant task. Plus, the target organizations are by definition the lowest of low-volume customers, and the cost of sales is high. As a result, the vendor community has largely fallen short thus far in penetrating this market.

Yet the reason to believe the vendors could hold a great part of the key to success lies in their mercenary motivation: profits. Free markets are wonderful enablers, and if there indeed is a latent need for technology in a large enough group, then the likes of Intel, Microsoft, IBM, HP, or Dell will be motivated to invest in it. On the other hand, we cannot expect that these multinational organizations would view their programs to be US-only initiatives, and there is no reason therefore to believe that US manufacturers would benefit any more than their German or Chinese counterparts.

Intersect360 Research Analysis

The advanced technology gap for SMEs is demonstrable and clearly linked to national competitiveness. Efforts to bridge this gap are worthwhile. To be successful, these efforts need to be well-informed, going beyond the notion of lower-cost hardware to measurably aid the manufacturer in the incorporation of CAE technology. Small manufacturers need lower cost, lower risk, and more expertise.

The HPC community is relying too much on the role of the federal government to solve this problem. It is a matter affecting the national economy, sure, but we cannot expect the government to mandate or to legislate digital manufacturing adoption.

Conversely, the potential role of academia is underestimated. Individual efforts such as the Blue Collar Computing initiative from Ohio State University or the STAR Program at the University of Texas have already had measures of success with commercial outreach, and in Montana the Rocky Mountain Supercomputing Center has begun following a public-private model as well.

Not all manufacturers will benefit from the adoption of digital manufacturing. Among those who would, the challenges of moving from a CAD environment to a CAE environment on the desktop are quite different from the challenges of moving from desktop to cluster. A successful initiative will be specifically targeted but broad in reach, and it must address the myriad needs of the small manufacturer. Only when the Middle is no longer Missing will we see for sure which initiatives worked the best.

Editor's note: The opinions in this article are the sole opinions of the author and do not necessarily represent the views of the Digital Manufacturing Report.