Leave it as it is. You cannot improve on it.
Leave it as it is. You cannot improve on it. — Theodore Roosevelt at the Grand Canyon, 1903.
That may be true of the Grand Canyon, but I am repurposing this quote to explain the hype that surrounds Industry 4.0 today. Many of my colleagues are defining the evolution from Industry 1.0 to 4.0 differently- but one major problem remains, they are trying to fix new problems with old solutions.
What I have seen from working in the industrial automation world is a market that is aware of the demands inside Industry 4.0, be that data analytics or cybersecurity — but with some noticeable ignorance. I have seen vendors adopting hardware and software combinations that are built to last decades with amazingly successful ecosystems built around their solutions (at least a few of them).
In this type of ecosystem developed over decades, there are many custom and one-to-one defined components. For example, products from PLC-Vendor-1 would only work with SCADA from SCADA-Vendor-1.
Let’s fast forward to the early 2000s with the rise of software tools. Everyone wanted to extend their hardware to many different types of software products — like MES or Historian. SCADA tools were generically designed and hungry for data coming from several vendors. Then there’s the challenging task to connect them all — every vendor’s PLC, DCS, Remote IO, Robots and, well, you name it.
Was normalized data under consideration? Nope. It was all about bringing data to a single vendor.
Was security under consideration? Nope. The best we have seen is “Basic Auth” to access a view of the application and eventually something with Windows AD.
Was threat protection under consideration? Nope. It was intended to be offline.
Was scalability under consideration? Nope. No one needed it.
Was remote management under consideration? Nope. Only a single site at first. (In engineering times, I used to combine multiple sites’ DBs to one replication to make it remotely managed).
Now enter OPC and a few more abstraction layers to solve normalized data access with an added security layer. Essentially, take any hardware vendor that can be used inside that OPC server and connect any software vendor that can consume OPC. It will solve that PLC to SCADA / Historian / DB problem seamlessly. So, the big question is, can you repurpose this 20 years of closed silo, custom assets-to-OPC or asset-to-SCADA system and open it for one-to-many Industry 4.0 demands?
The answer is no, do not leave it as it is. Theodore Roosevelt was talking about the Grand Canyon, not Industry 4.0. You cannot leave it as it is — you must improve on it and adapt.
Multiple vendors are adding “MQTT” and “AMQP” and calling it Industry 4.0 ready. That is surely a disaster waiting to happen — I have said it to every CIO possible. What changed all of the sudden? The whole dynamic of data hungry applications and how they can ingest the data in a wide array of formats with the medium (networks and cloud networks to be specific) to make that data available.
Cloud native apps, Edge native apps and everything in between started calculating KPIs of interest to customers. The physical medium from IoT gateways, networks and open-source cognitive layers became a catalyst to get those KPIs overnight. These call for an entire rework of products that are made for secure connectivity, runtime and analytics.
Can we change the plant and industrial systems inside it? Nope. Can we take that source of data and make it available in a secure, controlled and unified way? Absolutely.
I remember when our Litmus Automation team started discussing plans to redesign our “LoopCloud Gateway Agent” to “LoopEdge”. Our first concern was to always provide vast data (at that point in 2017, LoopEdge could already collect more data from PLCs, DCS and other assets than all those PLC-OPC products) in a controlled, optimized, and normalized way.
To accommodate it, we designed an edge computing product from the ground up by rewriting every protocol driver and communication method to put enhanced security, isolation and normalization in place.
The Industry 4.0 application developers, data scientists, and anyone with IDE in their laptops deserve better. Let’s not give them 90s systems in 2019. Do not — leave it as it is!
