Note, this article is continued from a previous instalement, available here.
There was lots of other good stuff during the Devil’s Advocate session earlier this month at Industrial 5G Forum. Notably, a couple of questions came up on regional spectrum policy, and the importance of so-called ‘vertical’ spectrum which is owned and controlled by enterprises themselves. David Rolfe at Volt Active Data reflected: “If you’re pushing use cases that are dependent on high volumes of data moving around reliably with predictable low-latencies, then the fewer external constraints you have… the better… You can’t talk about ultra-liability, and then not have an ultra reliable network. And if that means your own spectrum, then so be it.”
Of course, it is not possible to own spectrum in every country – unless you are a big mobile operator with deep pockets and millions of customers. Which is the case, as it stands, in India, noted Nitin Kumar at Bharti Airtel – although, equally, that does not mean enterprises cannot lease spectrum for private networks, or bundle it into a managed-service lease (which also provides public/private roaming). Kumar said: “If we talk about India, the spectrum lies with the telco operators only. There is no separate or vertical spectrum for enterprises… As of now, we are deploying all our private networks on Airtel spectrum.”
This commentary is expanded upon in the session, which is available on-demand here; other topics also came up, including discussion about mission-critical SLAs (also on low-power IoT networks), Wi-Fi trade-offs, and the superior role of 5G in the industrial tech mix. The last point was made by Kumar, invariably, and went like this: “Regarding private 5G, we are totally bullish on it. Going ahead in Industry 4.0, 5G will provide most of the answers to your problems. It will inter-work with different technologies… and offer better speeds, lower latencies, higher reliability… It has all the key ingredients to create a more connected, intelligent, efficient industrial sector.”
But the session overall – as written in the previous installment to this article already – took a more moderate tone, which better reflected the status and role of 5G within the enterprise market. The below bullets round-out our five-point summary of what the 5G bashers said about industrial 5G; the first two sections, listed below, are available here.
1 | HYPED AND HUMBLED – WITH RISK OF PERMANENT DAMAGE
2 | PART OF A BIGGER REVOLUTION – WHICH WILL TAKE DECADES
3 | NICHE TECHNOLOGY – WHICH EXPANDS INDUSTRY 4.0
The mood in the camp – that the hype is damaging, and that the revolution is bigger – did not stop the (rest of the) panel from making clear that 5G has a unique place in the tech mix. Yes, it is a niche one, currently, but it is also a novel one, which expands the whole Industry 4.0 discipline by enabling a range of low-latency and high-reliability wireless use cases. Responding to a question about which Industry 4.0 use cases are sprung exclusively (or most effectively) by 5G, Rolfe at Volt Active Data commented: “The one that really stands out is real-time video analytics, [especially when] tied to autonomous mobile robotics. And that’s come from a clear-sky perspective.”
He added: “We have customers doing this now… Instead of putting a really expensive graphics card and [AI] in your factory robot for $30,000 or $40,000 a go… you analyze [their movement] on CCTV and drive them by remote control. A human could [control] one of these, maybe, with a splitting headache after half an hour. [But] software can [manage it] spectacularly well – so long as you’re thinking about the collective … You can’t have 500 robots all pursuing their own goals. It would be a big scrum; like seven year olds playing soccer. So you need to impose authority and rules. And right now we’re seeing it in factories, but you could extend it to a whole campus.”
With that, Rolfe took the ball and ran – and imagined Tetris-like airports and car parks. He said: “You could quite plausibly establish a campus-wide 5G network [at an airport] and have robots drag planes to the end of the runway… or to busy urban car parks where you could have a bunch of robots playing Tetris with your car all day. Because they know when you are returning, and yours is five-deep at the back – and the robots would spend all day shuffling cars to save space, and eventually hand your keys back at the right time… These are the non-obvious but lucrative use cases we would see. But right now, the big thing is interpreting video and doing clever things with it.
4 | AN INDUSTRY IN TRANSITION – WITH LOTS TO DO
As much as the talk (hype) is about how 5G will change industry, the parallel trend is that 5G will change telecoms. Kumar at Bharti Airtel commented: “Telcos are the new tech-cos. They’re working with different spectrum [in different] industries. They’re in mining, transportation, manufacturing – all these sectors… We can provide groundbreaking solutions to industries, and we are actively working to build a robust ecosystem of partners – of technology providers, system integrators, industry experts. – to create a customized end-to-end solution for different industries.” But the cellular-IoT ecosystem is nascent – at the private 5G high-end, just as it is at the NB-IoT/LTE-M low-end.
No such problem for the LoRaWAN community, of course – which has been selling private networks to enterprises for donkeys’ years, and has a very well-established developer ecosystem to make IoT solutions-to-order. The LoRa Alliance has 400 members, noted Alper Yegin at Actility. He remarked: “The ecosystem is a key element. [For LoRaWAN] there are multiple other technologies that target the same use cases. Some came before [and] some came after. But, aside from the technical differences, the biggest difference is the ecosystem… [That is] what’s behind the exponential acceleration.”
The LTE market is well-oiled, but the 5G market has a long way to go, and will look enviably across the way (down its nose) at the old LoRaWAN brigade.
5 | RISING DATA STORM – TO BLOW THE DOORS OFF
Rolfe at Volt Active Data spoke very well about the bigger Industry 4.0 picture, which is how to manage the flood of data coming off new enterprise 5G networks – and, more urgently, how to monetize it in a way they can afford all the new edge infrastructure they need to manage it in the first place. He said: “Customers are seeing 10 times as much command and control data. Which is the first issue. The second is that 5G encourages stateless components at the edge – which creates more requirements for higher availability… [and] means we potentially have to support 10 times the data without 10 times the revenue to pay for it.”
He went on: “If you go up the value chain, the people using these networks may well face similar issues – because they’re going to have this big fire hose of data, and they’re not going to be able to keep it or store it [to] process it later because it will be valueless. So one of the bigger issues… is the economics and practicality of working with so much data. We understand it, and we have partners who understand it. But I am not sure the average developer gets it, or knows what they’re signing up for.”
Indeed, Rolfe went on to suggest a fundamental schism in the IT/OT worlds, which will not be so easily closed. He said: “There is a disconnect between current software engineering practice, which almost fetishizes change all the time, versus what a lot of people in industry want, which is to change nothing-ever once it is deployed… We are now playing in the real world with real machines – not virtual ones. And that requires a degree of discipline that I’m not sure the software industry has had before.”
