A few weeks ago I journeyed to lake country to visit with one of our best distributors. These folks are AB through and through. Their customers are solid AB supporters with deep ties to the company and the entire AB product line. During our wide ranging discussions, one of the topics was DeviceNet. We haven’t sold any DeviceNet for quite a while so I asked about how much DeviceNet they’re selling. Their answer was nearly nothing. They have a few legacy applications from years ago, but other than that they just don’t hear the word “DeviceNet” much at all.
It’s clear that DeviceNet’s time has passed. It was a good fifteen year run, but it’s over. Buried. Done. It now joins ControlNet and CompoNet in the industrial automation history books. DeviceNet at least had a much longer run than either of those technologies.
ControlNet had its time in the sun in the 1990s. Designed as the replacement for Data Highway, it offered an incredible (for the time) 5 Mbit speed, but was burdened with a pseudo-open media layer and really expensive ASIC, cabling, and installation costs that severely limited its viability. It found some applications with the highest, most well-heeled tier of AB customers that required the best performance. The wide availability of Ethernet killed it completely.
CompoNet, on the other hand, just never found market. Designed to compete against the more established and simpler to use AS-I bus, it never achieved any kind of market penetration. I can honestly say I never heard anyone describe their CompoNet application. I can’t recall ever getting a call from a device manufacture that wanted to implement it or seeing an actual live installation. There must have been some, but danged if I ever heard of any.
That means that three of the big four CIP protocols are no longer viable. CIP is, of course, the object oriented technology that provides the core of all four technologies. It is the object/attribute structure and the messaging (explicit and implicit) that’s used in all these protocols.
So what’s the future of CIP? Where does Allen-Bradley go with EtherNet/IP?
A few weeks ago I attended a presentation at 2nd Street in the old gym where the Bradleys used to watch employees in the Allen-Bradley basketball league. AB announced a rebranding of RsLinx to Factory Talk and the addition of OPC UA clients and servers to Factory Talk. Very nice. But in the question and answer portion of the presentation, they were specifically asked about the future of CIP (and implicitly in the question, EtherNet/IP).
Even though their presentation focused on OPC UA and the ways that they are integrating it into their product line, they expressed 100% support for EtherNet/IP and CIP. I think the words they used were “fully committed” and “CIP offers more than OPC UA.” I am not quoting exactly as the question was answered at the front of the room off-mike. The intention was clear. AB expects to continue using EtherNet/IP and CIP even though it appears that other technologies are surpassing it.
Personally, I think they are making a mistake. For many reasons, most of which I’ve written about, CIP and EtherNet/IP are nearing end of life. Faster networks with vastly more bandwidth, such as TSN (Time Sensitive Networking – deterministic Ethernet) and OPC UA Publish/Subscribe architecture is the network configuration of the future. I am hoping that internally they realize that and that they are saying what they are saying for political reasons. I certainly hope that their product managers see what’s happened to DeviceNet, CompoNet, and ControlNet and know that it’s just a matter of time until EtherNet/IP follows them.
One of the tenets of business management is that it’s better to obsolete yourself before someone else does. Personally, I love EtherNet/IP and the way it’s being used by AB controllers and hope that they find a way to keep it healthy, happy, and alive.
But I’m nervous about it.