One of the best-run conferences of the year is the ODVA Industry Conference. I have nothing but good things to say about it and the event organizers. It’s held every 18 months so it’s not so much an annual conference as a sort of semi-annual celebration of CIP, the Common Industrial Protocol, and its offspring: EtherNet/IP, CompoNet, DeviceNet and ControlNet.
As you might imagine, CompoNet, DeviceNet and ControlNet weren’t really mentioned except in passing. There just isn’t any real interest in expanding support for them, talking about them or even deploying them. Everything, as you know, has moved to Ethernet. DeviceNet still has a huge installed base, but as those machines are taken out of service, reworked or replaced, they are getting an upgrade to EtherNet/IP. Unless the newest CAN (Control Area networking) standard gains a significant foothold in the automotive market, we’ve probably seen the heyday of those technologies.
If you thought that everyone at the conference was talking about EtherNet/IP, well, you’d be wrong. EtherNet/IP is now pretty long in the tooth, having reached the ripe old age of 20 years last year. That’s pretty darn old for a technology, and it’s only due to the fact that it’s still the workhorse for the Allen-Bradly architecture that it still survives. But that doesn’t really give it it’s due. EtherNet/IP is an outstanding I/O protocol that is easy to use, operate and troubleshoot.
EtherNet/IP will continue to be one of the top I/O protocols for the next 20 years because it’s just good at what it does. The cyclic communication method for I/O data is exactly what is needed for automated systems on the factory floor. The acyclic communication method for configuration and troubleshooting data is exactly right for moving non-I/O data. EtherNet/IP has all the necessary features without being overburdened with functionality like PROFINET IO. It will be with us for a long, long time.
So, what was everyone at the conference talking about if not discussing EtherNet/IP? Most of the seminars were focused on alternative physical media for EtherNet/IP (and Ethernet in general). There is complete agreement that Ethernet is the technology for the foreseeable future. The question in everyone’s mind, though, is what medium that will carry EtherNet/IP and Ethernet in the next five or ten years?
There is any number of possible answers to that question including TSN (Time Sensitive Networking), some form of 802.11 wireless, 5G or other media. All the discussions are on alternative media technologies, not EtherNet/IP itself.
Noticeably, there was only some small discussion about 1Gig Ethernet. Why you might ask? Because 1 Gig provides little value to the automation engineer. When we are dealing with physical devices, there is only so fast that metal, plastic, paper or liquid can move, and none of those things can move fast enough to need 1 Gigabaud communications.
Personally, I gave a short talk on saving keys and certificates in a CIP Security system. My talk was well attended and generated a lot of discussions but, honestly, it was largely negative. In my research, I could not find a good way to protect the keys from intruders. Even the expensive TPM (Trusted Platform Module) devices have been proven to be unsafe in some cases. There is no good way that I could find to protect keys in a CIP Security system, and I said so in my talk.
Even key generation is a problem. Generating keys at your factory means that the key is in there and subject to intrusion at every step in your device’s journey to a customer: distributor, system integrator, machine builder, test facility…etc. A key can be stolen anywhere along the way and used later to infiltrate the end customer’s network. And if you generate it when deployed, you have the problem of using an algorithm and hardware random number generator that can create an adequate key. Lots of issues to solve around CIP Security.
All in all, it was a good, well-attended conference in excellent facilities. I am looking forward to attending the next one in the fall of 2021.