Advanced Computing in the Age of AI | Thursday, March 28, 2024

Cloud Computing and the Resurgence of Small- to Medium-Sized Manufacturers 

Cloud computing has many of the same goals as time-sharing, and has the added benefit of four decades of technological advances to support its mission of bringing affordable computing within reach of just about everyone. The drive to provide cloud computing services to the manufacturing sector was underscored by Fujitsu's announcement this week of its new cloud service.

Once upon a time, back when CIOs and top IT leaders were known as Data Processing Managers, there was something called time-sharing. It was a set-up that allowed the sharing of precious computer resources by a number of users through multi-tasking and multi-programming.

Here's the entry in Wikipedia: "By allowing a large number of users to interact concurrently with a single computer, time-sharing dramatically lowered the cost of providing computing capability, made it possible for individuals and organizations to use a computer without owning one, and promoted the interactive use of computers and the development of new interactive applications."

Sound familiar?

Fast-forward some 40-plus years. Cloud computing has many of the same goals, and has the added benefit of four decades of technological advances to support its mission of bringing affordable computing within reach of just about everyone.

The drive to provide cloud computing services to the manufacturing sector was underscored by Fujitsu's announcement this week of its new cloud service. The company's TC Cloud, which will be introduced in phases later this year, specifically targets manufacturing. Users will be able to run analytic simulations on the TC Cloud's Analytical Platform Service. Another service provides analytical applications and a help desk service will assist users in setting up and running those applications. Although the official rollout of TC Cloud has yet to happen, Nikon is already using the Analytical Platform Service.

Of course, Nikon is a very large company with extensive in-house HPC resources to support its use of the cloud's capabilities. And surely Fujitsu is betting that other major manufacturers around the globe will also find the new service attractive.

Attending to the SMMs

But what about the small- to medium-sized manufacturers — the SMMs who make up the majority of the manufacturing companies in the world? For example, there are about 300,000 manufacturers in the United States but only about five percent of them are big corporations like Boeing, Ford and Procter & Gamble.

As we noted in a June blog, the SMMs are a cautious bunch — they have some major reservations about becoming involved with high performance computing (HPC) and advanced digital manufacturing techniques such as modeling and simulation. Just to recap, here are some of the reasons why they are hesitant to embrace advanced digital manufacturing technology:

  • A reluctance to give up physical prototyping despite the cost — "It's not real unless you can hold lit in your hands."

  • Difficulty of using the software — FEA is tough, CFD is tougher.

  • Lack of talent — engineers skilled in using digital manufacturing tools usually head for the big manufacturers.

  • Integration — not only are expert computer scientists needed to work with these sophisticated tools, they must also interpret the results to other members of the manufacturing team.

  • Reluctance to take on the financial risk of adding modeling and simulation to existing processes, which might entail purchasing the required hardware and software and adding knowledgeable staff members.

With TC Cloud, Fujitsu will build an analytical simulation environment for its customers and then deliver the environment as a cloud service. The Analytical Platform Service can create a "hybrid environment" by combining the customer's computational resources with those of the Fujitsu cloud. The company claims that "large-volume analytic results can be conveniently navigated."

But if you're a small manufacturer and you have limited computational resources — no clusters or a supercomputers on a desktop to merge with the Fujitsu offering and a dearth of in-house HPC expertise — what then? Perhaps the Analytical Help Desk may be an answer, offering a variety of support services including running the analyses on behalf of the customer.

Concerns in the Cloud

Compounding the reluctance of the SMMs to engage in advanced modeling and simulation in general are misgivings about entrusting critical applications to a cloud-based HPC service. Security is a major concern, as is availability and performance.

For example, this summer a lightening strike in Dublin, Ireland, took down Amazon's cloud services in Europe for several days. A direct lightening strike on a power transformer that also disables backup generators is an unusual event, but it tends to make potential cloud customers a bit gun shy.

Other concerns include privacy, sudden and unexpected data loss, the ability to share data between different cloud services, and what happens to your data if you decide to cancel your current cloud provider.

Despite these misgivings, manufacturing in the cloud is gaining momentum. The Microsoft Discrete Manufacturing Cloud Computing Survey, which polled 152 IT and business decision makers in a variety of key manufacturing vertical industries in the United States, Germany and France, found that nearly half of the respondents replied that the biggest benefit of cloud computing is the lowered cost of optimizing infrastructure. Efficient collaboration across geographies and the ability to respond quickly to business demands were also cited by a large percentage of the respondents.

These same benefits, along with solutions to many of the SMM concerns cited above, are being addressed by a number of initiatives. For example, there is President Obama's Advanced Manufacturing Partnership announced last month, and the creation of Predictive Innovation Centers now in the gestation stage at the National Center for Manufacturing Science (NCMS).

Say Jon Riley, executive director, Design & Engineering Programs, NCMS, "PICs are reminiscent of the old time-sharing goals of bringing together hardware, software, and people in a collaborative model that makes the most of available computational resources. The PIC gathers the talent, the tools and the infrastructure to create a digital manufacturing storefront dedicated to helping the SMMs. It's a concept whose time has come."

The PIC program is just one of a number of initiatives underway to help SMMs that would like to make use of HPC-powered modeling and simulation but don't have the resources. Included are:

  • NDEMC (National Digital Engineering and Manufacturing Consortium) Midwest Pilot, U.S. Council on Competitiveness.

  • America COMPETES Reauthorization Act.

  • The NCMS Digital Manufacturing Strategic Interest Group (SIG).

  • Alliance for High Performance Digital Manufacturing.

  • Vendor-specific go-to-market programs.

Clouds Over Montana

It's a good bet that cloud computing solutions will figure in many of these programs. And, in some cases, it already is.

For example, the Rocky Mountain Supercomputer Center (RMSC) in Butte, Montana, is offering supercomputer services in the cloud to SMMs throughout the region and up into Canada. Its customers include companies — among them small- to medium-sized manufacturers — universities, government agencies, and tribal enterprises.

Cloud services are handled on the Center's "Blue Sky" supercomputer, a four teraflop (theoretical peak) machine backed up with GPU accelerators and plenty of memory. Visualization capabilities include remote and immersive 3D, especially valuable for manufacturers tasked with modeling difficult geometries. One of RMSC's most valuable services is the support it provides to its clients to make maximum use of the Center's computing cloud.

The NCMS PICs and RMCS are a blueprint for the future. The availability of low cost, pay-as-you-go, secure HPC in the cloud, backed up by much needed hand holding, will allow SMMS to make the move to advanced digital manufacturing and reap the benefits.

It's a good bet that the cloud will play an important role in revitalizing the manufacturing sector and prove to be a particular boon to smaller manufacturers seeking to leverage advanced modeling, simulation and analytic computational capabilities.

EnterpriseAI