This article continues from here. It is an excerpt from a full report, called ‘Crossing the IT/OT divide – from co-creation to co-configuration, and how to bring industrial IoT to scale’. The report, part of Enterprise IoT Insights’ ongoing Digital Factory Solutions series, is available to download (for free) from here. Other entries in the Digital Factory Solutions series can be found here.
Let’s pause, and play the blame game, even if momentarily. Is there a sense these two factions are entrenched and guarding their territories? Is it one side’s fault? Is one one side blocking progress, as the other agitates for it? “I mean, yes,” says Martin at ABI. But he hesitates, initially, to apportion blame (he is more explicit, later).
Others, mostly on the OT side, are happier to do so. “There is no magic box,” says Bry Dillon, vice president of cloud, channels and community at OSIsoft in a robust attack on IT interlopers in the OT space. He is on stage, speaking to a captive audience at PI World in Gothenburg, back in September, as the event host.
“You don’t just throw data up into one big place, and put some algorithms on it – and hope it gives you insights for operations.”
OSIsoft’s PI System, it might be noted, has existed for decades, since before the idea of slamming together OT into IT to create IoT took hold. Sixty five per cent of Fortune 500 companies use it, the marketing types proclaim; a number of these industrialists are in the room in Gothenburg. Accenture suggests at the same even that 60 per cent of industrial IoT ‘platforms’ use PI System as their source file system.
Dillon makes clear the industrial set is indispensable in the face of new industrial IoT providers, promising automation and intelligence. He runs through a series of anecdotes, like he is reassuring old friends in a bar. We have time for just one, here. “I was sitting with a chemical company, and the gentleman across the table starts to chuckle,” he says, relaying a conversation from an ‘executive briefing’.
“I asked why he was laughing; he says: ‘We had a consultancy in here, claiming ‘the brightest data scientists’ – that if we provided our data, they’d answer questions we didn’t even know to ask. So we brought them in, and three months later we kicked them out again. [Because] we can’t teach these guys engineering – they don’t know the difference between a pump and a heat exchanger’.”
Dillon addresses the audience directly. “These people aren’t just any people,” he says, and the room hunches forward. “They are operating the foundation of our entire economy. It’s these people, when empowered with the right operational data, that have the ability to not only transform their world, but to transform our greater world.”
KPMG acknowledges the issues, but suggests the IT-side is improving, and actively seeking collaboration. “Tech organizations are pretty good at engaging domain experts,” says Greg Corlis, managing director for emerging technologies at the consultancy. “In some cases there may be an air of arrogance on [their] side, but they tend to listen to the domain experts.”
He adds: “It has changed over time. [They used to] produce solutions they felt would meet the business need, but often missed the mark. This resulted in expensive development cycles and a lack of adoption. [But] over the last few years they have started to see the value of co-collaborating with industry experts.”
He also defends, correctly, the premise of IoT platforms as a means to solve interoperability hitches in IoT deployments, which remain a feature of new technology projects.
“Organizations are constantly approached by tech providers offering point solutions… and left to solve the problem themselves, and acquire platforms designed to aggregate across IoT technologies. Once an appropriate platform is available in the corporation, plugging in new IoT devices is relatively simple and only requires a small effort to integrate those products,” he says.
But we should seek out the middle ground. The OT crowd’s back is up, anyway, it seems. “Many of the initial forays into instrumentation and bringing data out from the shop floor were driven by the IT world. This caused a degree of resentment from the OT side,” comments Ian Hughes, senior analyst at 451 Research.
At the same time, the IT market recognizes it cannot go alone; OT know-how is needed. “Marketing for industrial IoT has changed from an IT centric delivery of an IoT platform to a more complete OT-focused solution to a business problem, which may include elements such as IoT platforms and machine learning. Start-ups that have significant OT experience tend to get more engagement than pure IT ones,” says Hughes.
PROBLEM STATEMENT
So where are we up to, then? 451 Research has polled the enterprise market about collaboration between IT and OT on IoT projects, and offers a qualified (albeit limited) view of around 500 industry types. Its 2019 ‘voice of the enterprise’ says a little under half (45 per cent) of enterprises reckon cooperation between the sides is close, “from concept to operation”, and a third (34 per cent) says cooperation is just about sufficient.
But as many as one in five admit a failure to bridge the divide, citing either “active conflict” about project control (15 per cent), or else no working relationship at all (five per cent). Nevertheless, the sense is, from the poll, the job functions are merging; nine in 10 (87 per cent) said IT and OT roles will either blend in brand new job types (46 per cent), or steadily overlap (41 per cent).
Martin at ABI says the joining of IT and OT in IoT should be redefined. “This idea of IT/OT convergence is a problem statement. The challenge is around integration and collaboration, and not convergence. Because that’s the goal. That’s what the market is trying to achieve, ultimately.”
Integration and collaboration are the twin principles of co-creation, of course. The point, we should remember, is real change will not come with point solutions; the idea of industrial revolution, by definition, seeks to make enterprises anew, and raise them up as more than the sum of their parts. Spot purchases get through, notes Martin. He gives a couple of examples.
Dutch firm Realwear, maker of industrial augmented reality (AR) headsets, has gained traction not just because of the ruggedized nature of its HMT-1 product, but because of the way it has positioned it to IT departments – as a “voice operated Android computer” to be “worn on the head”. Martin comments: “IT has budget for computers; it doesn’t have budget for smart glasses.”
Similarly, he points to the success of the Ubimax system, the popular back-end software for industrial AR applications, as a standard-issue floor in certain enterprises. “DHL staff can order a pair of smart glasses just as they would order a smartphone, to support workflow. It used to be they’d have to go through all these proofs. Now they don’t.”
But point solutions do not bring enterprise-wide change. The ambition of Industry 4.0 is to go beyond pick-me-up line applications to redefine entire operations, where data insights are gleaned across factories and factory footprints. “Quite right,” says Martin. “That’s the real sausage making, and it’s highly complex and highly volatile.”
He explains: “First-generation automation was around low line-level point solutions. Now it’s much broader. It’s about the entire facility, and not just a single line. It’s also hinged on newer metrics – things like operational equipment effectiveness (OEE). And to achieve that, you need sight of everything in your asset field and the ability to allocate capacity based on availability. That takes so much other stuff.”
The process of co-creation, between IT and OT specialists, seeks to explode business-wide opportunities by industry type, but also for each player within each industrial discipline. Again, it does not work without input from the OT cohort. A minute’s downtime in the automotive sector is worth about $30,000, reckons Martin; an additional minute’s productivity is worth the same.
“The question is about the cost of not embracing Industry 4.0? When OT folks see this, it’s very clear. But they need to involve IT to make things happen. That’s where things start to break down. That’s where we talk about pilot purgatory, and this whole IT/OT convergence issue comes about.”
And there is the blame. One easy way to get around it, he suggests, is to install an IT liaison within the OT team – an IT manager to work on a project basis on an OT deployment. “That’s how you facilitate collaboration and the exchange of ideas early on, and stop things from dying in the pilot phase,” suggests Martin.
This article continues here. The full report, part of Enterprise IoT Insights’ ongoing Digital Factory Solutions series, is available to download for free from here. Other entries in the Digital Factory Solutions series can be found here.
