Future of the Factory Floor


I’m hearing a lot of talk about IT protocols and the factory floor. This stems from the fact that the Factory Floor is slowly becoming IT. It’s much like the continents drifting around the globe over time. Continents form, move together, move apart over thousands of years. Continents of IT and the Factory Floor are now moving together.

It’s going to happen a lot faster than some vendors in factory automation would like. It’s also apparent to me that IT is going to absorb the factory floor and make the Engineers who work on the factory floor part of IT. As that happens, I think vendors will respond by making their products have more IT “friendly” features.

If you look at IT kinds of products they use a lot of XML, SOAP, SNMP and Web Services. Things pretty easily integrate. You can find interfaces available on the network, figure out what they do and connect and use those interfaces pretty easily. There are things like WSDL documents that publicly describe what is available. We all know that it’s a lot harder on the factory floor. We have things like EtherNet/IP, DeviceNet, Profibus DP and PLCs.

What’s this evolution going to look like? Well, certainly it’s clear that PLCs will look a lot more like Servers with built-in routers over the next few years. They’ll support Oracle and Sequel databases, HTML5 Web pages and web services. You’ll probably have the ability to write control code in Java or IEC 6-1131. We won’t lose all the factory floor protocols – those things will still do all the hard work of communicating with I/O.

What will be different is that these PLC Servers will be easily integrated with the Enterprise. They’ll look to the Enterprise just like any other Enterprise Server. It’s just that there will be a back end to these devices that does factory floor I/O and Control. With the speed and bandwidth of Servers today and the ability to partition the hardware and software it will be easy to use standard Servers as PLCs in the future. In fact, there is no reason to have the physical footprint of the PLCs we have today. A hardened server will do the trick nicely. I expect that Siemens, Rockwell and other PLC manufacturers will probably become software companies in the future at least in regard to the control currently found in their PLC’s.

The key, of course, is going to be security. That’s where OPC UA comes in. I expect UA to be the prime Transport layer for all the Enterprise / PLC communications in the future. It has the ability to support IT transports like Web Services as well as fast binary. With its adaptable, user configurable and powerful security component it makes the most sense as the prime data mover between the factory floor and Enterprise.

I do expect HMIs to disappear also. Everyone will have their own person HMI on them with their own web pages or dashboard that they created for their own personal use. Why have a fixed HMI with mostly fixed screens that are a pain to change? That’s not going to work in the future. As soon as you get close to a machine, your pad or phone will sense that and pop up the dashboard or browser screen for that machine.

So the sensors and actuators won’t change? No, not at all. I can see these devices also having OPC UA and while still communicating to the PLC Server machine using traditional I/O protocols they will also be sending other information to the Enterprise. They might send energy data, maintenance data, diagnostics and what not to the Enterprise directly. With almost every microprocessor having an Ethernet MAC there is no reasons to believe that this isn’t likely.