There are few terms in the history of technology that are so misused, so misunderstood and possibly overhyped as the Internet of Things (IoT), sometimes called the Industrial Internet of Things (IIoT). It’s often regarded as something new but as for the plant floor, manufacturers have always worked to get data from manufacturing systems and integrate it with their business systems. The ability to move data into the enterprise improves the ability of manufacturers to improve efficiency, enhance product quality and lower costs.
With factory floor IoT being so pervasive, with an overwhelming number of options and sometimes-confusing components, a number of myths have evolved around it that need to be explored and discarded:
Even though a lot of people think IoT is about moving data, it’s not! IoT is about making more intelligent business decisions. That isn’t necessarily true in the consumer world where the objective for an IoT system can be to entertain or improve the quality of a person’s life; but in manufacturing, we need to make more intelligent use of resources, be more efficient, nimbler, faster and more responsive to our customers. That’s what the factory floor Internet of Things is really all about.
No, technology isn’t even near the top of the list of key aspects of IoT. That was true 25 years ago when all we had was serial buses like Modbus RTU, DeviceNet, and Profibus DP. But today, the key to IoT is a viable business model; knowing how you can use data – what’s available and what you might possibly get – to enhance your key performance indicators. It should be just like any other improvement you make to the factory floor: how much are we going to spend (more than you think) and how will we achieve a return on that investment.
It’s actually not; there are limits to what we can do with today’s infrastructure. Once we start equipping every motor, actuator and sensor with IoT capabilities, there won’t be enough storage for all that accumulated data. We’d have to cover the planet with server farms to store it all! The IoT infrastructure needs to evolve to the point where data can be automatically curated and analyzed locally with only summary data transmitted to offsite storage. AWS and Microsoft Azure and the rest are only the first generation of the IoT infrastructure that our manufacturing plants will need.
The Raspberry Pi is great for quick demonstration projects but as a long term, supportable technology on the manufacturing floor, it’s much too limited. The Raspberry Pi is a very underpowered IoT device, unable to perform the curating, analysis and summarizing of data that will be needed in tomorrow’s IoT projects. It’s the equivalent of equipping a new motor drive with an 8-bit processor.
This may never get sorted out. IoT may be destined to become fragmented with oodles of competing standards. The market is too big, the applications too numerous, the vendors too disparate to ever coalesce around a group of standards, let alone a single standard. There are many other technologies. One technology, named Kafka, is even named after a Russian poet!
And if you’re a device manufacturer making drives, robots or even simpler devices, the landscape is going to be especially chaotic and complex. You will need to think about what customers you’ll serve and how you’ll use devices, services, applications, hardware platforms, standards groups, business models, geographies and even other users to serve them. It’s going to be a lot harder than just adding an EtherNet/IP or ProfiNet IO to your device.
These are both myths and truths at the same time, depending on your perspective. IoT is easy to do if you just want to move some data and you aren’t really concerned about security, meta data, longevity or Return on Investment (ROI). Just moving data is easy (see Myth #10).
IoT is hard to do if you want to do it right. That means securing the data, executing a project that completes on schedule, using hardware and software components that are viable for the long term, and making money (ROI).
There are advantages and disadvantages to using open source in the manufacturing world. Technologies like OPC UA are extraordinarily feature rich and extremely complicated to deploy. There can be significant risk in going with an open source project in a complicated manufacturing environment where millions of dollars can be at stake. If you choose to use open source, find a reliable partner, fund your own internal expert or ensure that the open source project is viable both today and, in the future, and will be extended as the technology evolves.
The ODVA and the PI (ProfiNet International) trade associations propagated this in the early years of IoT, but it’s finally been put to rest. Early on, both the ODVA and the PI trade associations tried to pass off EtherNet/IP and ProfNet IO as “cloud ready,” implying that you could connect them to the cloud and use them in IoT applications. ProfiNet IO and EtherNet/IP are excellent technologies for moving I/O data between field devices in a factory and a programmable controller. Nothing more and nothing less.
Not even close on this one. If you’ve been in manufacturing long enough to break for lunch, you should know we’ve been moving machine data to back-office systems since the first programmable controller was deployed. And if you’re thinking this is new in commercial applications, have you heard of something called an automated teller machine (ATM)? What is an ATM but a device for moving data from remote locations to a central server? Seems like an Internet of Things device, doesn’t it?
What is new, is that we now have much better enabling technologies for the Internet of Things than we’ve ever had in the past.
MQTT is very popular and deployed in a multitude of applications. But a viable IoT technology in 2022 and beyond requires support for standard data modeling. MQTT, in its Sparkplug B extension, supports its own type of data modeling. This may change in the future but as of this writing, it has no ability to support the kinds of standardized data modeling that many trade associations are promoting. You can connect MQTT to Cloud applications, but it is anything but seamless.
But MQTT can be a good choice as well for many small IoT applications where data modeling is not currently important. It has a large number of adherents because of its small footprint, simplicity and “push” architecture. It is a reasonable solution for IoT prototyping, small, in-house applications and other applications not requiring widely adopted data models. It may not be the right solution for the coming era of open, widely available data models for all automation devices.
For manufacturers that want to future proof their applications, OPC UA might be a better solution. OPC UA is the leading choice for Smart Manufacturing applications who anticipate a future where data modeling is key to productivity and competitiveness. OPC UA is the only solution that promises support for all of the data model architectures being developed for Smart Manufacturing systems of the future.
MQTT and OPC UA are both viable for IoT applications but there is no perfect or right solution! Like many other technology decisions in automation over the last 20 or 30 years, there is not and will not be, a clear winner. Both technologies will grow and prosper. Developers, integrators and system architects will eventually come to learn where they have the best success with OPC UA and the best success with MQTT. Manufacturers are going to have to understand, support and work with both technologies and determine which works best in which applications.
Real Time Automation is uniquely positioned to support manufacturers in this environment. Our communication gateways support both OPC UA and MQTT technologies. That means that control engineers and operations staff now have the ability to reach into a Siemens S7, an Allen-Bradley Logix controller, an Allen-Bradley PLC5 and other controllers and devices to extract information and publish it to an MQTT broker or present it as an OPC UA Server.
Like all RTA gateways, the unit is extremely configurable. You can pick single tags, User-Defined Tags (AB), Data Space Blocks (Siemens) and other data using any one of the many protocol drivers in the RTA protocol suite. And on the MQTT side, the gateway publishes that data using topics that you define to the brokers you identify.
Our soon to be released factory floor data collection software platform will also support both MQTT and OPC UA and collect data from many different factory floor technologies, manipulate it, scale it, normalize it and send it out over OPC UA and MQTT..
John S. Rinaldi is Chief Strategist and Director of WOW! for Real Time Automation (RTA) in Pewaukee, WI. With a focus on simplicity, support, expert consulting and tailoring for specific customer applications, RTA is meeting customer needs in applications worldwide. John is not only a recognized expert in industrial networks and an automation strategist but a speaker, blogger, the author of over 100 articles on industrial networking and the author of six books including: