By TJ McDermott
Vernor Vinge's 1986 book "Marooned in Realtime" is a murder-mystery/time-travel novel. The time travel in his book is one way into the future. Those who travel beyond the 23rd century discover that humanity has disappeared from the Earth. The survivors believe that the exponential rate of change in technology allowed humanity to transcend to a different level of existence.
This concept has stuck with me throughout my professional career. I see this ever-accelerating rate of technological change right now in the ever-growing population of different process automation protocols available to industrial automation—Modbus TCP, Profinet IO, EtherNet/IP, Foundation fieldbus, EtherCat, VARAN, SERCOS-III, FL-Net and Ethernet Powerlink—just to name a few, and just in Ethernet. I'm looking at a list of 35 different protocols generated by at least 16 different companies or organizations.
Sure, sure, you can point out the benefits of each one of the automation protocols, and why each was developed to fill a need in the field. The fact remains that they seem to propagate like rats, and we're threatened with being overwhelmed trying to keep up with them.
A company for which I used to work is rolling out its next-generation machine. Before changing employers, I had a chance to contribute wishes and opinions toward its features. The generation prior to this new one used Mechatrolink. For this new generation, I expected to see Profibus, or maybe CANopen or SERCOS-III. Nope. When the new generation was unveiled, I was told it used CompoNet. I had to do some research to find out what that was.
Is it so bad to want to have a single standard? Is it so bad to want the pace to slow down just a bit? I see the term Luddite being aimed at my head, but can't you see Ned Ludd's point? In the case of industrial Ethernet, we've not come close to the limitations of speed and flexibility such that we need an increasing number of different flavors. Even Microsoft doesn't foist new office document formats on the world as fast as network protocols seem to pop up.
The answer to any "why?" question pertinent to human interaction will always be "money." Why are there so many protocols? Different companies or organizations saw a way to profit while solving a perceived problem. I wish these groups would focus on improving one standard, rather than starting their own.
I'll bet large companies do as well. Large companies with which I've dealt have standards for the software they use. They pick a software version, and suppliers must use that version in the equipment provided. The selected version could be two or more revisions older than current, but it means there is some stability within that company's software suite.
I like this approach. I've mentioned previously about the frustrations of finding software bugs crossing major revisions. A selected, stable-version approach eliminates this to some extent, but also means that these companies get left behind in the expanding rate of change of networking. If a large end user picks version 17 or 2.9 or XP as a standard, then they won't be able to use that new network protocol that leaves all others in the dirt, and can talk to unlimited numbers of network nodes in real time.
On the other hand, maybe I do deserve the term Luddite. After all, we did run out of IPv4 addresses in February 2011. IPv6 addressing takes care of the number of addresses problem. But how many industrial networks are prepared for IPv6? I don't see many industrial devices advertised with IPv6 capability. Even the companies pushing their favorite new protocol aren't keeping up.
On the gripping hand, maybe I don't deserve the title after all. I'm not asking to go back to the good old days of my favorite protocol. I don't see my livelihood threatened as the Luddites did. I'm simply frustrated at the senseless proliferation of protocols. I'm all for faster speeds and increased capabilities, but we've not yet hit hard limits; Gigabit Ethernet hasn't hit the industrial floor in a big way yet.
I might be put off by the proliferation of protocols, but I'll deal with it. Just as long as they don't create one called Skynet.
TJ McDermott is senior project engineer at Systems Interface (www.systems-interface.com), a control systems integrator based in Bothell, Wash.
