ringsThere is a rather startling shift in thinking about how to architect control systems. For many years now, we’ve use a much more hierarchical model for automation systems. I’m sure in your career you’ve seen this automation model more than once.

It’s the hierarchical automation model that every plant in North America uses for its enterprise systems. The new thinking is that the model should be compressed. A LOT! In fact, it looks like the new model is for everything to be connected to everything in a single very fast, high bandwidth Ethernet ring network.

It’s very sophisticated with sub rings and multiple segments. The future is an architecture where anything can talk to anything. Everything is on the same IT/OT backbone. Sensors, cell phones, motor drives, servers, HMIs, printers, your desktop, the chiller, the fire alarm and entry systems, and everything else on the same backbone.

The core of the physical layer for a system like this is TSN – Time Sensitive Networking. TSN is based on a bunch of standards that, if you’re like me, you try to ignore. Things like 802.1AS (time synchronization), 802.1AR (device identity), 802.1AE (MAC security), and more. Sometimes I think the folks at Campbells Soup Company are on the IEEE committee. These standards specify how the switches are going to cooperate to make sure that the deterministic stuff (controllers talking to motors) gets done reliably and deterministically while still getting the audio, video, and everyday internet search traffic (checking basketball scores) though reasonably on time too.

Even though I’m an electrical engineer, I’m more interested in the software end of this. How are all these devices going to communicate reliably on this network? What sort of protocols make sense for a network where everything talks to everything? And how do we secure this network?

If you’ve followed me for a while, you know I’m a proponent of OPC UA. I try to be honest in my assessments. Here’s what makes me think that OPC UA will be the prime driver for a network like this:

1) UA has the built-in security to secure devices on this network. And because end users get to decide which devices get secured and how they get secured (tokens, passwords, certificates), they have a lot of control. They like that. None of the other protocols have that kind of built-in security.

2) UA has a Publisher–Subscriber organization. I think that’s the future, not the cyclic operation we use today. Having devices like Motors and Valves publish data when needed makes sense to me. Cyclic operation just doesn’t make sense for many devices we use on the factory floor. We’ve always had to pervert EtherNet/IP and ProfiNet IO to support devices with intermittent data like barcode readers, scales, and the rest over those networks.

3) UA has the most sophisticated addressing and data model of any technology I’ve seen. It’s adaptable to anything and that data model can support all the metadata that IT applications can use to make sense of the data.

4) The Pub-Sub extension (publisher to many subscribers) not to be confused with Publisher–Subscriber (publisher and one subscriber) makes UA adaptable for the IoT and cloud applications.

This is all very new and there is still a lot of work to be done, but I’m looking forward to using all this new technology in the future.

If you’re new to OPC UA, you should get my blue OPC UA book. It’s the most easy-to-read and comprehensive book on OPC UA there is. You can find out more about it by clicking here: OPC UA comprehensive guide.