10 April 2001

A new view from outside the cave

by Perry Sink Marshall

In September I visited the IMTS (Machine Tool) show in Chicago with my colleague, Rick Goldstein. Rick worked in that industry for more than 15 years. He pointed out to me how much smaller the machine tool controls portion of the show is now than it was years ago, due to the fact that a controller that cost $50,000 in 1985 may cost $5,000 or less now and have many more features. When hardware prices drop 90%, it tends to shrink your market.

Similarly, several months ago in this column [November 2000], I wrote about Napster’s threat in the music industry and the parallel problem that exists in automation: Networks, faster processors, and declining hardware costs are driving profits and revenues down for automation companies. It’s true that the controls business is slowly shrinking. The $1,000 PC on your desk has 100 times the power of the $2,000 PLC in your factory.

What have automation companies done to combat this? The fieldbus wars of the 1990s were a strategy: introducing fear, uncertainty, and doubt to anything that threatens to commoditize the hardware. After all, if everyone uses a network, then each device becomes "just another node," which can substitute for something similar from another company. Now companies have to compete on value instead of having exclusivity. If they can’t prevent people from using networks, they have to at least persuade them to use one that favors their products more than someone else’s.


Now let’s consider an extreme scenario. Suppose starting 1 July 2001, all controller, PLC, and DCS hardware cost $1,000-that’s right, anything with a CPU, memory, storage, and a network card for a mere $1,000.

And let’s further suppose that all the software is absolutely free. Microsoft Factory is included as a standard part of Windows 2001. Ladder logic, PID programs, and operator interface software packages are tossed in free when you buy TurboTax, along with Acrobat Reader, Internet Explorer, and 700 hours of free AOL. What would happen to the controls business?

Well, the first thing that would happen is that Monster.com, Jobs.com, and Headhunter.net would have a flood of resumes from PLC and HMI salespeople. But the customers would do just fine. Integrators would, too, as long as they didn’t tie up their uniqueness in knowing someone’s proprietary control platform.

Integrators whose true skill is deep understanding of a particular type of process would do very well, and they would simply earn their money by adapting commodity hardware and software to specialized applications. Customers would spend less time sorting through the smoke and mirrors of compatibility issues and spend more time solving the problem at hand. Imagine that.

One peculiar thing about our business is that it’s always two to three years behind the rest of the world. If this makes things a wee bit boring at times, the nice part is that you can always get a job as a soothsayer: Simply walk outside your cave, take a look at the "real" world, and go back inside. Offer predictions to your fellow cave dwellers based on what’s already happening outside.

Well, I’ve been walking around outside my cave, and guess what? The scenario I’ve just described–$1,000 controllers running on free software–has already happened. I’m referring to the huge Web server market, where a third of all systems run on free software (Linux) and where even if you do spend money, you get a ton of extras thrown in.

The information technology world is quite accustomed to this. In fact nearly all of IBM’s applications run on Linux, many of those e-commerce IBM commercials you see are advertising Linux solutions, and IBM has committed more than $300 million to developing additional Linux services over the next three years. It’s quite a headline: "$87 billion company builds huge business on free software."

Blue chip

This is all to say that IBM has become a "blue chip systems integrator" that sells billions of dollars of custom solutions based on commodity hardware and free software. No doubt part of its success comes from talking to customers about their business process instead of debating the merits of hardware platforms.

After all, how complex does an automation device really have to be? Last I checked, most control systems were still running Boolean logic and either analog or digital control loops. This stuff ain’t rocket science. Any senior electrical engineering student can fully grasp the concepts. Why should a PLC cost up to $10,000?

I’ve heard dire predictions about various control vendors and the diminishing prospects of selling 10-year-old proprietary technology at inflated prices. I haven’t been hearing the same thing about IBM. And with the same forces at work in our own industry–industrial Ethernet, open source software, and the growing complexity of manufacturing processes–the future is very, very bright for those who can take inexpensive tools and expertly apply them to real customer problems.

After all, isn’t applying expertise to real problems why we came here in the first place? IT