What is Industrial Data Ops?

Dec. 26, 2023
With system integrator and technology supplier insights, we take a look at the rise of industrial data ops—how it differs from historical data storage and trending activities and how it can be used to predict future trends to optimize supply chains and derive strategic business insights. 

TRANSCRIPT

Welcome to this Industry Update episode of the Automation World Gets Your Questions 
Answered podcast. I’m David Greenfield, editor in chief at Automation World. These 
Industry Updates are a new addition to this podcast series as a way to help keep our 
listeners updated on the latest goings on in the automation industry. 

For this Industry Update, we’re focusing on industrial data ops – what it is and what you 
need to know about it.

So, while the phrase about data being the new oil has definitely been a bit worn out by now, 
it still remains true. After all, everything about Industry 4.0, digital transformation, the 
Industrial Internet of Things, digital twins and artificial intelligence—they all come down to 
data aggregation, contextualization, transmission and visualization. While this sounds 
simple enough since all the devices and equipment on the factory floor generate tons of 
data, the catch lies in making sense out of all this disparate data and compiling it in a way 
that it can be used for a company’s benefit.

That’s why you’re hearing more and more about data ops and industrial data ops more 
specifically.

Daniel Malyszko of system integrator Malisko Engineering explains that industrial data ops 
is about bringing together system architectures, data access methods and cultural 
approaches as well in the way industries manage and use their data. He says it's about 
creating synergies by adding context and governance to data, and ensuring that various 
data sources across the enterprise merge fluidly to drive optimized results with shortened 
data analytics cycles.

Now, of course, to deliver on this grand concept, industrial data ops require a common data 
platform that allows for data modeling, contextualization and data consumption across 
multiple applications, and to be able to do this at scale.

According to Daniel, industry’s realm of focus has historically consisted of integrating PLCs 
and HMIs to deliver immediate access to process data for analysis by site maintenance, 
engineering, production, quality and continuous improvement personnel. But as industries 
are becoming more digitally interconnected, simply storing and trending historical data 
isn't enough. 

As others in the company want access to this data for their own decision making purposes, 
we face more intricate questions, such as: 
*    Where and how is the data directed post-historization?
*    Who is leveraging this data and for what purpose?
*    What other data sources can add valuable context to the existing data?
*    How do various applications use the data? 
*    How is the data secured and governed?

Analysts at the enterprise level are increasingly advocating for a cloud-first approach to 
data management because of its scalability, flexibility and its compatibility with an array of 
big data toolsets. For many corporate users of plant floor data, it's not just about 
understanding what happened in the past, Daniel says, it’s also about predicting future 
trends to optimize supply chains and derive strategic business insights. 

To provide an example of how this is playing out in industry today, Daniel referenced a 
recent project Malisko Engineering was involved in that would seamlessly bring together 
data from production, quality analysis, research and development, and engineering. On 
paper, he says this project seemed pretty straightforward, but it involved straddling 
multiple data worlds, each with their own intricacies. 

In this project, Daniel said they noticed that production used a cloud-based MES system, 
they combined spreadsheets from quality and R&D, and they used a homegrown laboratory 
information management system and historian data.  Meanwhile, the engineering 
department used historian data with HMI trending tools, but they also wanted to overlay 
lab data, specifically manual data entries from the laboratory information system. 

To harmonize these disparate data sources, Malisko Engineering crafted a common data 
model independent of the data’s origins. Daniel said one of the most valuable aspects to this 
deployment was introducing event data, such as production stages, steps and alarms and 
adding context to time-series data. By doing this, each department can maintain their own 
distinct tools, but the enterprise benefits from a unified and contextualized data landscape 
for more efficient analysis and decision making. 

Looking back on this project and at how industrial data ops is transforming industry, 
Daniel said it's no longer just about the controls engineering department’s choreography of 
data but orchestrating what he calls a grand symphony of data across the company. 
Looking at this from a higher level point of view to understand how industrial data ops is 
being applied across industry, John Harrington of HighByte, a supplier of industrial data 
ops software, says that standardizing device-level data into structures is key to industrial 
data ops, but it’s really only just the beginning. And he highlights four key reasons why 
manufacturers need an industrial data ops system as part of their information architecture. 

His four reasons are these:

Number 1, You’re dealing with machine and vendor variability.  John says that because 
standards bodies are made up of vendors and users, as the standard is being defined, 
variances are allowed for vendor systems with unique capabilities and limitations as well 
as for specific use cases. While the intent of this is flexibility, of course, the result is often 
ambiguity because it’s typical for vendors to implement the same standard slightly 
differently. Also, vendors tend to refine their systems and change data models over time to 
suit their needs.  This means that even minor variations in datasets require human 
interaction to link these machines to other systems in the network to automate dashboards 
or analytics. Industrial data ops technology helps users connect to a range of sources 
without having to write or maintain code. That allows for non-standardized input to then 
be modeled and transformed to the governed data standard for any use case.   

Number 2, without industrial data ops, John says you’re viewing individual data without 
any relationship context. When analyzing operational metrics, bottlenecks or quality root 
causes for a production line, specific information from each machine, test stand and sensor 
should ideally be assembled into a single payload for that line.  Industrial data ops makes 
this possible he says by merging and modeling data from multiple machines and then 
correlating those data by logical use case. 

Number 3, John says you’re looking at more than just device data. You can’t make strategic 
decisions if you’re not linking your machine data to other systems across your 
organization. This includes enterprise applications, such as your ERP system, and your 
manufacturing databases, such as SCADA, MES, Historians, and maintenance management 
systems. Industrial data ops should be able to connect to virtually any system in your 
organization and combine information from these systems with machine data to bridge the 
gap between operations technology and the rest of the business.  
  
And number 4 involves not getting the data you need when you need it. As well all know by 
now, information overload is a real problem in the Industry 4.0 world. So, understanding 
what’s needed—and when its needed—is critical. Sometimes data is needed at a cyclic rate, 
say, once per second. Other times, you may need an event-based feed to identify cell 
production completion, defects or machine performance issues. John says industrial data 
ops technology can define the desired data payload and its event or frequency to create a 
more efficient decision-making process and minimize your cloud costs because, in this way, 
you’re only storing and processing the data you need. 

If you’re interested in learning more about industrial data ops and the data hub and spoke 
model, be sure to visit this podcast on the Automation World site. There, just above the 
transcript, I’ve included links to several recent articles on these topics.

So thanks for checking out this Industry Update episode of the Automation World Gets Your 
Questions Answered podcast series. And remember to keep watching this space to stay on 
top of the latest news, trends and insights on the world of industrial automation.