I read with interest Willy Shih’s (Professor of Management Practice at Harvard) and Helmuth Ludwig’s (EVP at Siemens PLM Software) article in the Harvard Business Review on the Biggest Challenges of Data-Driven Manufacturing. In it, they discuss the idea that Data-Driven Manufacturing and the widespread deployment of low-cost sensors connected to the Internet (the so-called Internet of Things or IoT) will be the next wave driving efficient and responsive production systems, but that we’re in a somewhat hyped stage and managers will need to understand some of the key challenges associated with the paradigm shifts.
I couldn’t agree more, but I wanted to refine and somewhat shift some of the challenges they’ve identified. As exciting as the IoT is, I am not sure it is really the main driver to the next stage of “efficient and responsive” production. Or at the least, I’m not sure it’s the path to grow the efficiency and business of most our manufacturers and job shops related to CNC machining.
Shih and Ludwig identify the following challenges:
- A Move from Time-Triggered to Event-Triggered Control Systems
- A Need for a Unified Data Model that delivers Data Sharing, Not Just Data Exchange
- The Challenge of Integrating Legacy Systems
- Security Challenges once we connect everything to the Internet
It’s the first three areas coupled with the question of whether we should let the oh-so-trendy IoT tail wag the dog that I want to talk about. We’ll talk a bit about that fourth area of security too, but in a somewhat different light that will be more obvious. I want to drill down on each topic in turn so we can keep continuity and have a discussion around the original author’s ideas.
Moving From Time-Triggered to Event-Triggered Control Systems
I read this and my immediate reaction is to wonder whatever happened to Lean Manufacturing? Wouldn’t a key definition of Lean Manufacturing’s precepts be that Manufacturing be Event-Triggered and not Time-Triggered?
From this perspective, I welcome organizations coming to an Event-Triggered perspective, but I would encourage them not to require IoT as the catalyst. Lean Manufacturing has made sense long before IoT was even a thing (not clear it is even today quite yet a thing). It’s not a tremendously easy transition to make, but it is a transition that many have made, that is well understood, and that we can find many resources to help us with. It’s costs and benefits are readily quantifiable based on the work of others.
We don’t even need a truckload of expensive new software nor shiny IoT gadgets to make it happen. Toyota and many others since managed to get there with humble Kanban Cards and other techniques. If you’re not familiar with these things, check out CNCCookbook’s multi-part article series that introduces you to Lean Manufacturing. It won’t take long to get oriented to these ideas and understand why they’ve been so valuable to so many for so long.
Perhaps there really is nothing much new under the sun.
Unified Data Model that delivers Data Sharing, Not Just Data Exchange
Most product design and manufacturing operations have benefited extensively from computerization: from ERP systems and computer-aided design, to engineering analysis and simulation programs that enable virtual prototyping, to manufacturing execution systems and automation design, all the way down to robotic automation systems on the factory floor and in materials handling. Most investments over the last three decades have gone into point solutions for design and manufacturing, and the integration of the “transaction oriented” ERP systems. Once a design was finalized, manufacturing used the engineering bill of materials and manually added relevant manufacturing (manufacturing bill of materials) and process data (process bill of materials). But a major challenge arises from the fact that because these systems were designed independently, they weren’t designed originally to talk to each other.
Yes! That! I will paraphrase the quote:
We built point solutions around specific Design Office and Shop Floor problems. At the same time we built ERP solutions that were “transaction oriented”. In other words, they were mostly useful to the Bean Counters and didn’t really serve the needs of the actual Manufacturing Process.
More modules, or just better data sharing among modules we already have?
But that leads to a bigger question: What is needed to serve the needs of the larger Manufacturing Process? What will turbocharge the optimization and efficiency of that process? The authors seem to argue that the problem is things like propagating changes to Bill of Materials upstream. If only we could tie all these point solutions together, we would have the answer. If only we could share data, is their watchword.
I want to offer a dissenting opinion. Sharing data is all fine and well, but it will only make somewhat more efficient the processes we already have in place. It is therefore an evolutionary step. It’s not clear to me that making data sharing easier has ever resulted in Revolutionary Progress throughout the course of Enterprise Software.
Let me say that again, and more emphatically:
It is not clear to me that sharing data has ever resulted in Revolutionary Progress for Enterprise Software.
It may be that Manufacturing is so fundamentally different that data sharing will revolutionize it, but I don’t think so, having been around multiple Enterprise Software tracks as well as Manufacturing.
Here’s what does consistently revolutionize the Enterprise: entirely new application categories that simulate, analyze, and automate workflows that were never addressed by existing applications.
I submit that we have a huge opportunity to do that facing us in Manufacturing. We’ve barely scratched the surface, in fact. If we look at perhaps he most mature ecosystems of application categories in Enterprise Software today, there are 3 categories:
- Accounting, including ERP
- Sales and Marketing, including CRM
Everything else, including Manufacturing, is nowhere close to being as mature in terms of delivering complete, off-the-shelf, and sophisticated Best Practices solutions to anyone that wants to pay for a solution across all categories from the Global 2000 to small business.
The Integration of Legacy Systems
If only we could integrate our Legacy Systems…
When all you have is a hammer, everything is a nail. I see this argument about the challenge of Integrating Legacy Systems as stemming largely from the perspective that our biggest opportunity comes from Data Sharing. That implies all the necessary systems are in place and that Legacy Integrations of systems never designed for graceful Data Sharing is a big challenge.
But, if you agree with my view that we have yet to even build the software for the vast majority of the processes, then Legacy Systems are not a problem because there aren’t any. The Legacy Systems we must replace and integrate with are pencil and paper ad hoc systems. They’re emails and phone calls between Design Offices and Shop Floors.
In short, when we build this next generation of applications, we will have the opportunity to take stock of integration opportunities and build them in from the get-go.
Security: Always a Challenge to be Reckoned With!
If progress is predicated on IoT–thousands of inexpensive sensors and other devices all connected to the Internet, security is going to be a boondoggle, there is no question. But how much will the IoT really drive out future in Manufacturing? Do we really foresee attaching these sensors to raw aluminum and titanium so it can be tracked as it is placed on CNC machines for work? Is that really going to make us any more efficient?
OK, how about putting said CNC machines themselves into service as smarter devices that can comminicate sensor results as needed? Is there value in doing that over the Internet, or are we just as happy using dedicated wired and wireless networks inside the Shop Floor? I submit the latter is plenty good enough and we’re already well in control of how to manage that approach. We may want to up our security further there, but that’s okay.
As far as the wave of new applications I suggest, those will most likely be Cloud applications–the world isn’t financing new development of much else any more. As a result, it’s fair to consider the security ramifications there. But this is an entirely different security challenge than the one of “extremely fine-grained control and monitoring” mentioned in the Harvard article. I’m not that sure we need to tackle the latter at all.
Organizations today are used to centralizing access to the Internet and shielding it with the appropriate Firewalls and other safeguards. Inside the Firewall, physical security is important. This is again well understood and does not represent some new paradigm or challenge that has to be rediscovered. We need merely decide this is the model we want for our IoT activities and apply the existing protocols appropriately.
The same is true for that next generation of applications that will be Cloud-Based.
This is a two-part series. In Part 2, I want to talk about that vision for entirely new applications. What might they be? I have some very concrete ideas I want to share and get your feedback on. Stay tuned for the second installment. If you haven’t already subscribed to our weekly newsletter, be sure you don’t miss out. There is a sign up form directly below and joining it will give you not only the weekly newsletter, but access to some really special content that only our most Loyal Readers have access to.
Like what you read on CNCCookbook?
Join 100,000+ CNC'ers! Get our latest blog posts delivered straight to your email inbox once a week for free. Plus, we’ll give you access to some great CNC reference materials including:
- Our Big List of over 200 CNC Tips and Techniques
- Our Free GCode Programming Basics Course
- And more!
Just enter your name and email address below:
100% Privacy: We will never Spam you!