OPC UA – Love at First Sight

In 2011, I discovered OPC UA. It was love at first sight. I was totally enamored nearly forgetting my first love, EtherNet/IP. I thought the theory behind it was brilliant. I thought it was going to be consequential for manufacturing. I thought it would revolutionize the shop floor. I was so enamored that I wrote two books on it (OPC UA – Unified Architecture: The Everyman’s Guide to OPC UA and OPC UA: The Basics).

I was sort of wrong.

OPC UA is many of those things in Germany. The Germans have an advantage over us. The government, industry and universities all get together and pick a path. Then they all follow it. The path they picked is OPC UA and they have followed it to the letter. To sell in Germany, Europe for that matter, you really need it. It’s that pervasive.

None of that is true in the US. In the US, we don’t get together and decide anything. We fight it out in the marketplace and the best marketed technology wins (Note that I said best marketed technology not best technology). We are cowboys here. Everyone has a better idea of how to do things. That’s why there are arguments about EtherNet/IP, PROFINET and Modbus TCP. It’s one of the reasons why the US is so innovative. Both systems have their advantages and, if I had to pick, I like more innovation and less rule following, but that’s another article.

OPC UA has some commercial and technical issues that have prevented it from having the success I had hoped for it. Let’s look at the technical issues first. The massive, complex and unwieldly specification is one of them. It’s hard to describe what OPC UA is because it’s SO many things. I shudder when a customer says they want OPC UA. They (and I) can’t easily define what that means.

My biggest complaint is that the vendors behind OPC UA want it to solve every problem. They want it to be a low-level device protocol, the cloud protocol, the IT communication system, the peer PLC communications system. There wasn’t an application they didn’t want to address. When you do that, you end up with a system that doesn’t meet the requirements for any application well. It’s excellent for a few things, OK for lots of things, mediocre at other things (it’s IoT solution, for example).

On the commercial side, major US vendors have really stood in the way of OPC UA in the US for a long time. These vendors had nothing to gain from using and promoting it. They waited until they built out competing systems that they could sell against it. They didn’t disparage OPC UA, they just didn’t push it. I think they saw it as just another way for competitors to get access to their data and manufacturing architectures. With so much of the US market, tied to a few vendors and those vendors not all that supportive, OPC UA hasn’t been a huge part of the conversation in the US the way it is in Germany.

I still think OPC UA has a place in Smart Manufacturing. It does some things very well. It models data very well. You can use it to standardize data models before sending it to the enterprise or the Cloud. But in some cases, it’s easier and faster to build a system using MQTT. MQTT doesn’t do 10% of what OPC UA does, but what it does do, it’s cheap, fast and easy to use.

Sometimes the love of your life just doesn’t work out. I hoped for a lot more from this relationship, but didn’t get it. But there’s always hope for the future.