The problem with the industrial internet of things (IIoT)

If you do even a cursory search about the IIoT you'll be overwhelmed within minutes. Every company of any respectable size has an "IIoT Platform" that's going to totally change the world is already saving them trillions of dollars and teaching machines how to think and dance and write poetry. The marketing is sleek, the message is vague, and volume is high.

After a few pages of search you'll start to see a common dissent, security. It's a real risk, and one that needs active work now, but that's not what I see as the biggest issue with it.

As Chief Scientist at Predikto, I've seen a lot of these IIoT initiatives from the inside.  Hundreds of companies, all around the world, titans in their own industry, are seeing the hype and marketing around the IIoT and they're hungry for it. They have unplanned downtime, they have scheduling delays, their supply chains aren't perfectly efficient, the use cases are straightforward, understandable, and they're tied to enormous amounts of value.  The most advanced companies in this group are ready and are starting to see value. They're many years into the journey and have invested in sensors, communication networks, data consolidation and tagging, and basic reporting.  Most I think have the cart well before the horse.

First, a story of two perspectives.  Enterprise asset management is a system for managing the lifecycle of some asset, in this case, a "thing". A company purchases a thing, installs it, uses it, maintains it, and eventually decommissions it, and all of this is tracked in some EAM system to provide a single source of truth to different business units.  The industrial IoT was supposed to be an internet of these things, the things we spend money on and care about.  As implemented, in the wild today, I mostly see an internet of sensors, not things.

So why does that matter, aren't those close enough? Not really.  To really understand the past, present and future condition of an asset, as the IIoT aims to, the semantics the system must be thing centric, just like the EAM systems that came before.  The signals coming off of sensors must be tied to the global asset hierarchy, just like work orders and other EAM information.  And yet, most implementations have sensor streams and fault codes tied only to the very top level of the hierarchy, or worse, only a unique sensor identifier is present, flattening the hierarchy entirely.

The problem with the IIoT, as I see it, is one of relevance.  I only care about this thing, what is relevant to it? What is relevant to it one level higher in the hierarchy? Or below? Good implementations can answer that, and huge value can be derived from those datasets, but globally that's a small elite group.



Will has a background in Mechanical Engineering from Auburn, but mostly just writes software now. He was the first employee at Predikto, and is currently building out the premiere platform for predictive maintenance in heavy industry there as Chief Scientist. When not working on that, he is generally working on something related to python, data science or cycling.

Leave a Reply