Ronald Regan once said that the nine most terrifying words in the English language are “I’m from the government and I’m here to help”. We could bring that into the Manufacturing Industry by simply changing it to “I’m from IT and I’m here to help”.
Having spent more than 30 years in Information Technology (IT) with 20 of which associated with Manufacturing and Utilities companies, I am acutely aware that IT has a bit of a reputation with Operational Technology (OT). With IT being seen as overbearing, not particularly serious and prone to error.
IT will occasionally attempt to involve itself with manufacturing and operations, typically, grossly underestimating the complexity of the task and attempting to foist solutions upon them. Bringing the same tools and approaches that they use for marketing and finance to an area with significantly different requirements.
In a recent presentation to the Smart Manufacturing Week Expo at the Birmingham NEC I posed a few questions to the audience.
1 – Does IT understand our challenges?
2 – Does IT understand our technology?
3 – Do you collaborate with IT wherever possible?
Rather like a small child in a classroom desperate to be noticed, one audience member stretched his hand as high as possible and glanced around furtively trying to encourage others to raise their hands also. One out of roughly one hundred. That feels about right.
This tear-down of IT is all very entertaining but is avoiding the deeper issue. What is the cost of this disconnect? In my opinion, it’s huge! Successful organisations are able to blend their manufacturing with their enterprise data. The problem is, OT just can’t do it alone. Only IT has the tools to gather data from marketing, finance, etc. It just wouldn’t make sense for OT to run what is an enterprise-wide resource.
By bringing all the relevant data into one easily accessible repository, tasks that would have required significant effort can become simple. Some typical examples include reporting CO2 production by line, scheduling maintenance that can include planned and remaining holidays for staff, actual cost/profitability of individual machines/lines to name a few.
So, what’s the issue? Gathering OT data means gathering telemetry, historical and operational systems data, which is generally locked in an OT Network and defended by gatekeepers who, based on previous experience, do not want IT anywhere near their stuff. Oh, and typically IT doesn’t have the first clue about industrial technologies such as SCADA, Modbus, OPC/UA, BACnet, MQTT, TimeSeries, Digital Twins etc.
Enter the “Modern Data Platform”. For several years organisations have been moving towards Cloud Based Data Platforms. At the core of these platforms is the Data Lake, bringing the benefits of huge scale with storage being as cheap as $10 per TB per Month. This is something of a game-changer. So, we’ve now got a platform to store all this information, how do we get it in?
I’m going to introduce some architectural patterns here: “Medallion Architecture” describes how data can be ingested in its raw form, validated, conformed, cleansed and transformed ready for use, “Lambda Architecture” works with “Medallion Architecture” to introduce the concept of a “Speed Layer” for real-time data, streaming in from live systems. So, we’ve now got a pattern for handling bringing in large volumes of fast-moving data. How do we connect to the Manufacturing and Operations Data?
Well that’s the tricky bit. The good news here is that the big players in the cloud industry Microsoft/Amazon/Google have all been furiously trying to solve this problem for years now as it represents a big market opportunity for them. They are providing services that can make this all possible, including Data Collectors, Gateways etc. Each sufficiently secure to be allowed to see within the OT network.
Having collected, cleansed and prepared the relevant data, you are then in a position to leverage some of the really powerful tooling that is now available. e.g. Machine Learning, Anomaly Detection and Self Service Reporting.
With a suitably architected Modern Data Platform, the data becomes a resource for the entire organisation allowing OT to not only prepare and consume their own reporting, but also develop their own Machine Learning Models etc.
To conclude, in the authors opinion, the next step for a large number of manufacturing organisations should be to collaborate with IT to drive out further efficiencies and improve decision-making by pooling their data resource.
For further information on how to create a shared data platform, please contact Dufrain.
Stephen Armory, Head of Architecture at Dufrain, delivers innovative solutions across insurance, financial services, and manufacturing. Specialising in data warehousing, big data, and cloud technologies, he excels at conceptualising and implementing strategic data architectures that drive business value and guide organisations through technological transformation.
For more articles like this, visit our Industrial Data & AI channel.