Stephen Mellor picks up where he left off, and where we left him: the challenge with industrial IoT to gain widespread adoption is as much to do with interpretation as with deployment. Different industrial disciplines have different demands, even if they sometimes appear the same. Semantics come into play, he explains.
“The language is different. Manufacturing, say, has a whole vocabulary that doesn’t apply in healthcare.” He is in the process of explaining the logic for the frameworks and guides the Industrial Internet Consortium (IIC) has set down, and the array of ‘test beds’ and test drives’ it has set in motion.
These are designed to help join IT and OT for enterprises, and accelerate adoption of industrial IoT. The clearest example of this disjunct in understanding, as digital technologies go into different industrial sectors, is with the terms ‘safety’ and ‘security’, he explains.
“Safety, for example, means different things to different people. You and I understand it in a common way. But its meaning is entirely different in a hospital, say, compared to a factory floor, which is again entirely different to an electricity grid.” Patient safety, worker safety and public safety are variously at stake, in extreme cases.
A step removed, technology and operations are at cross purposes about security too, with the one dealing with cyber attacks and the other with physical attacks. “A factory manager thinks of security as locked doors and guards; the IT guy is thinking worms and firewalls. It’s the same word,” says Mellor.
The end goal, to disrupt business or rob property, may be the same, but the crime fighting is different.
It amuses him, he says, that both these words – safety and security – are the same in German (‘sicherheit’). Industrial dialects must be understood, and translated, if the technology industry is to speak the same language, consistently, with the new enterprise markets it is trying to serve.
“The challenge is to bring these things to each market in language they understand,” says Mellor. The IIC has just released a new volume of its technical dictionary to help. It offers simple definitions; version 2.2 includes a revised description of ‘digital twin’, new entries for ‘physical entity’ and ‘physical entity of interest’, the difference between ‘sensor’ and ‘actuator’.
As with language, and as with technology, the linguistics of industrial IoT are alive; ‘virtual entity’ has been eliminated in favour of ‘digital representation’. Fluency is required if technologists are to find a middle ground with industrialists. “You have to be able to talk their language,” says Mellor.
There are other important discrepancies between industrial disciplines, too, which need to be understood and articulated by the technology market. He returns, again, to security. “Everybody knows we need to be more secure. But, you know, do you put your trash in a bank vault? No, you don’t; you chuck it on the street.”
Balancing risk and cost, which spirals upwards in both directions in the case of security, is a delicate calculation. It is being made practical as industrial IoT technologies take hold, but the market is led by technologists, still. “It still isn’t coming from a burning business need – no one’s saying, ‘I have to do this, help me make this happen’,” he comments.
There are architectural challenges, too. Manufacturing is hierarchical – “a factory is a bunch of production lines, with a bunch of machines, each going “chk, chk, chk’.” A self-contained industrial system can be easily transposed in trees and subtrees onto a spreadsheet, and into a data lake for an analytics engine. But structures are different elsewhere. “Try to do that with solar energy, which is completely distributed. It works in a plant; it doesn’t work outside. The same architecture won’t work over here, and vice versa.“
The IIC was founded in 2014 to accelerate the development of the industrial IoT market. “There’s a trigger, although this kind of technology has been around for a long time. But at some point, people get excited about it.” The likes of Bosch, Dell, General Electric, Huawei, and Microsoft are founding and contributing members.
Its framework documents have sought to define the converged IT/OT landscape, offering broad-brushed reference designs and guides for data and systems integration, and of course security. It has since published a number of white papers on best practices, taking its frameworks into execution scenarios. But these are only baby steps for new industrial discipline that will take years to deliver on its promise – and will likely mutate and grow bigger than we can imagine.
The future is unknowable, says Mellor. “We recognise this is the beginning.” The first uniform resource locator (URL) to identify and locate web pages was invented in 1994, as a functional means to share documents. “Look at what we use it for: I mean, like, everything. So we don’t know where this is going.”
The IIC is seeking to order the basic principles for this new industrial world; the bigger questions are for the rest of society, and are not even being asked yet, he suggests. “Just consider: we’ve had the retirement age bumped in the UK by five years overnight – nobody’s thinking about that. What does that even mean in an age of robots? Do we have enough robots? Do we have too many? There are all sorts of questions, and we have no idea what the answers are.”
What’s Mellor’s background? “I’m a geek,” he says. An academic geek or a corporate geek? “Somebody who likes to get a salary.” He chuckles: “I like that.” He has a whole career behind him already, in real-time and embedded systems; he worked at CERN, moved to Berkeley and Stanford, set up his own company (“modelling languages”), which eventually sold to Mentor Graphics (now owned by Siemens).
He has been engaged in “standards work” for the Object Management Group, a computer industry consortium and parent of the IIC. Retirement is somewhere on his radar, it would not be impolite to say. But history helps to tell the future. How do we get to a point where complex domain expertise, entrenched through three industrial ages, is understood by technologists?
How do we get to a point where industrial IoT solutions are mapped to every dark corner of industry, so they can be lifted and scaled, and the fourth industrial revolution explodes as foretold? “I wish I knew. But what happens is it bubbles up,” says Mellor.
You can’t tell the future, it seems; you can only set the groundwork, so it comes more quickly. That is what experience says. That is what the IIC is seeking to achieve, except the market is moving beyond frameworks and dictionary definitions, at last.
“We need to go from proving technology to proving value,” he says, borrowing a line from the market. The IIC’s ‘test beds’ and ‘test drives’ are geared towards more practical know-how, to drive adoption. “That has to be the focus now – not frameworks, not best practices, but adoption. End of story.”
Its test-bed programme has been around for five years already, and is actually a tech exercise, to prove the gear works. It caters to the dream of industrial IoT, conjured by technologists, and not to the reality, defined by the industrialists. The new test drives seek to put this balance right, make the technology relevant, and answer the question about value.
It announced three last month, to show the value of ‘connected worker safety’, ‘intelligent video’, and ‘IoT sensor implementation’. More will come, says Mellor. “We’re not trying to prove the technology anymore. We’re looking to deploy it.”
He says: “We don’t talk about putting sensors on workers. There is a change in language. You have to have the technology, but then you need to say connected worker safety, or intelligent video – or whatever the heck it said in the press release.”
It has a third initiative, too, which joins with its test beds and test drives as part of its ‘accelerator’ project; a ‘challenge’ programme, which seeks, instead, to corral technologists around real business problems, presented by industry. The problem is the start-point, rather than the solution; the idea is the technology is not force-fitted into place.
The problem is clearly defined and sent to contestants, which do their thing; six months later, whatever, a jury makes an award for the best approach. At this level, we’re still not hacking out code so much as saying, ‘well, this way works’,” explains Mellor.
A pilot follows, hosted by the enterprise that initiated the challenge. The prize money, put up by the tech providers, funds the live tests. The IIC is running challenges for smart buildings, smart logistics, and smart construction. More will follow.
“You have to be able to show it in the hospital, which is different from the factory, which is different from the solar panel array.”
The point is that the tech industry has to keep its eyes on the road, to stay on a future path. There is no point, really, gazing the mists on the horizon. The IIC is contributing, in its measured way, to the market’s scatter-gun pattern of trial and development. Like he says, the future bubbles up.
“We’re trying to construct multiple innovation vehicles. We don’t care what others are doing. Everything is valid. We’re just trying to demonstrate that this stuff works for industry.”
