Wanted: New Ways to Model Information in Automation

DataModelingIn lots of ways, when you compare the technology used in Industrial Automation to what’s happened to IT over the last five, 10 or 20 years, IA has remained pretty unsophisticated. The reasons for that are well understood and they’re good reasons: large capital costs for equipment, long-term operational requirements, the costs incurred to upgrade technology, and all the rest.

But that’s changed a lot in recent years, and it’s going to change more. There are new requirements now, and to meet those new requirements, how we model information has to change. We can’t meet the new requirements for forward integration with our customers, backward integration with our vendors, and integration with our enterprise systems without better mechanisms for modeling information.

This is a priority for trade organizations, such as MDIS http://www.mdis-network.com serving the undersea oil and gas industry, BACnet http://www.bacnetinternational.org/ serving the building automation industry, and the RFID Association. The only way they can rein in integration costs and project lead times, which are increasing in all their industries, is to:

•Standardize access to devices in their industry

•Distribute those standardized models in a cohesive way

•Implement those models in a structured way

•Reference those models so that devices know exactly what information is available and how to access it.

So, What’s an Information Model?

A logical representation applied to a physical process, an Information Model can represent something as tiny as a screw, a component of a pump, or something as complex and large as an entire filling machine. The Information Model is simply a structure that defines the component devoid of any information on how process variables or data within that structure can be accessed.

All our standard networking technologies use Information Models to some extent:

EtherNet/IP and DeviceNet – These technologies use a model based on objects composed of attributes. The model is very limited. There is no mechanism to create hierarchies or relate objects to each other. These technologies allow Information Modeling in the most minimal sense.

ProfiNet IO and Profibus DP – These technologies are not much better. Both use a representation based on emulating an I/O rack with modules, slots, and points.

Modbus RTU and Modbus TCP – These technologies may be the worst at Information Modeling. Data is limited to registers and coils, with no organizing structure on that at all.

OPC UA Alone Standardizes Information Modeling

OPC UA – OPC UA is the best at Information Modeling because it’s been designed from the ground up to process information and not data. What’s really exciting about Information Modeling in OPC UA is that OPC UA is the only technology that has standardized the documentation, implementation, reference, and access to Information Models. OPC UA allows end users, trade associations, vendors, and others to create, distribute, and implement Information Models in a very organized, detailed, and structured way.

And even more important, the Information Model has nothing to do with how that information is stored or made available. In an OPC UA Address Space, there are specific mechanisms to browse that address space, access the nodes that comprise it, and encode, secure, and transport the information it contains. All of that is important, but in OPC UA those mechanisms are distinct and separate from how the information is modeled.

Free Book on OPC UA
Learn more about Information Modeling in my new book on OPC UA. From now until April 1, you can sign up for a free advance copy by contacting me here: FREE OPC UA BOOK. Simply send me the message “OPC UA ME!”