Industry 4.0 Q&A: “Manufacturers already tried LTE and it failed miserably,” says KPMG
How will incoming cellular technologies transform manufacturing, and spur its drive for smarter production? That is the question Enterprise IoT Insights put to both sides of the industrial divide in a new editorial report, called ‘Industrial LTE and 5G: How Incoming Cellular Technologies Will Transform Smart Manufacturing’.
Both factory operators and telecoms vendors were polled on the matter, as part of the new report, the first from Enterprise IoT Insights in a new Digital Factory Solutions series. As well, the article incorporates commentary from independent commentators, including advisory firm KPMG, which judged the broad claims of the telecoms sector that LTE will serve factories as ‘baloney’.
LTE, said KPMG, will work as a stop-gap for non-critical factory communications, but that is it. The role of cellular networks as essential components of industry’s new digital infrastructure, and the vaulting ambition of the telecoms community to reinvent itself for a new market, rests almost entirely on 5G, it seems. It was an interesting discussion.
Here is the full transcript of the interview with the team at KPMG, including Greg Corlis, the firm’s IoT leader, Michael Flaherty, its director of technology enablement, and Eric Logan, its principal for corporate strategy. Again, here is the link to the full report, to see the same discussion play out among a wider audience.
How do you see the role of cellular in the manufacturing space today?
Corlis: “Cellular is quite limited within manufacturing today. It’s been tried and tested, all the way from 2G through to LTE. The issue is there’s a lot of interference in these large manufacturing plants – in terms of noise, heat, friction, you name it – which really corrupts the transmission on the cellular network.
“So most factories these days – probably 90-plus per cent – are still running on ethernet, with some wi-fi on the floor, although wi-fi is susceptible to the same interference issues. With 5G, you can deploy small cells and MIMOs to the factory floor, and eliminate some of these challenges with beamforming. It will be more effective from a communications standpoint.
“But the technology is really new. We saw Bosch’s demo at Hannover Messe, which wasn’t impressive. The devices just aren’t there to attach to the machines – which is why you have these 5G gateways being attached. But the expectation is 5G will eliminate a lot of the challenges for wi-fi and other cellular technologies today.”
Logan: “Manufacturing success is inherently about minimising variability – in supply, in manufacturing, in demand. Technology is intended to give data access to make decisions to assist in this purpose. In many cases, rolling out a technology that answers only part of the variability challenge tends to create more or new forms of variability, which inhibits the success of the manufacturing process.
“While there has been some use of cellular technology, it hasn’t been widely adopted because manufacturers have not seen the level of reliability necessary to risk introducing additional variability.”
The suggestion, from many, is you can go a long way with LTE, actually – that 80-90 per cent of industrial use cases can be served with LTE. Does this issue of interference hamper LTE, just the same? Is there any appetite for LTE, or is the industrial market waiting for 5G?
Corlis: “Manufacturers tell us they’ve tried LTE, and it has failed miserably. So the market is waiting for 5G. You can get closer to the equipment than you could with LTE, at least without lots of extra work. The clients we’ve spoken with, anyway, are not going to do anything with LTE; they will wait for 5G, and what 5G brings to the market.
“Vendors like Nokia are focused on LTE first, evolving to 5G – in any type of venue, really. And that might work in some scenarios, but it would be very challenging in the manufacturing space. In warehouses, say, LTE is more than fine.
“But these other [interference] issues remain a challenge in manufacturing, which is why a lot of manufacturers are holding their breath for 5G before they make any dramatic moves to adopt true cellular solutions in factories.”
Flaherty: “But private LTE is maturing just as 5G comes on the horizon, and LTE supports other types of connectivity, as well, like NB-IoT and LTE-M. The thing is 5G, in general, brings so many more advanatges, in terms of latency, slicing, reliability.”
Are you saying, ultimately, that LTE won’t cut it on the production line, and at the ‘coalface’ of critical industry functions?
Corlis: “Not necessarily, it depends on the particular use case. Some manufacturers will end up deploying private LTE but more are inclinded to wait for 5G based on the additional technological benefits it provides.One of the biggest challenges is that 5G sensors are just not available yet.
“We expect to see a significant ramp up in 5G devices in mid to late 2020. That said, manufactures can deploy 5G networks now to replace wi-fi and begin to replace ethernet networks while they wait for 5G sensors. There are still advantages that can be realised for these manufacturers with this approach.
Flaherty: “It is almost too easy to say, ‘Yes, LTE can do that’. The actual transmission capabilities of many protocols could provide a solution on that basis alone. But there is a more holistic response that accounts for a wide range and diversity of use cases and network type communications that an enterprise wants to deploy today and into the future.
“Also let’s not forget the next generation of technologies that live on top of the network. With this broader lens applied, 5G demonstrates clear advantages over LTE.”
Logan: “I return to the goal of minimising variability, and many manufacturers believe the broad applicability and breadth of use cases for 5G networks will potentially increase reliability enough that it will outweigh a shorter-term use-case-by-use-case implementation of LTE technology.
“Furthermore, many are waiting because they want to ensure the most technologically advanced and broadly applicable network before they make any significant investment.”
So, to be clear, the message from the manufacturing fraternity is that LTE at coalface is just not good enough, and 5G is the only solution they will consider?
Flaherty: “It’s not that black and white. It comes down to use cases. LTE has some viability, but 5G just has more utility in every aspect.”
Corlis: “Agreed; it’s not so black and white. There are some use cases that LTE serves very well. But certain scenarios, which demand high bandwidth and low latency, will not be served by LTE. Of the manufacturers we talk to, some have limited use of LTE, but none use cellular for their broad factory operations. 5G will help to converge LTE, WiFi and Wired networks into single 5G networking solution.”
If certain non-critical use cases, however limited, are well served by LTE, why are we only talking about private LTE now? Why has manufacturing waited until now? Why has telecoms waited until now to address this sector?
Corlis: “Some carriers tried to go the private LTE route and garner interest from manufacturing and other sectors, but it was hard to demonstrate the value proposition. Manufacturers, already have wired networks so their needs to be a significant change agent to instigate a change in technology.
“Manufactures that are either building out a new factory or re-tooling a production line are prime to look at 5G as a solution to reduce costs and provide flexibility. I don’t see many manufacturers opting for private LTE over 5G, but other sectors may jump on a private LTE solution.
“That’s one of the big reasons private LTE hasn’t been quickly adopted, because wi-fi and ethernet can serve most purposes – so LTE offers less value in manufacturing environments. But 5G is different, because it brings an opportunity to rip out and eliminate ethernet on the factory floor. That’s where 5G starts to have appeal. LTE just doesn’t offer the same throughput as an ethernet solution.”
Can we talk about the 5G roadmap, in terms of standalone (SA) and non-standalone (NSA) 5G, and the development of it in three phases – of eMBB, mMTC, and URLLC? Will factories wait until Release 16 version in 2021/22, which introduces these key URLLC capabilities, or will they deploy asap, starting with eMBB / NSA 5G?
Corlis: “They will deploy when they can. Some use cases will adopt 5G technologies sooner rather than later, with the 5G gateways demonstrated at Hannover Messe. Especially if they are re-tooling and rebuilding plants.”
Flaherty: “They will deploy when they can, and the conditions make sense. Will Release 16 compliant equipment really take until 2022? It is certainly plausible you will see releases beforehand. Will an industrial plant wait to deploy 5G SA?
“If an industrial plant only has a need for technology solutions that are sensitive to extreme latency and strict timing requirements, then they would most likely wait for Release 16 or until the latency metrics they need are met.
“However, the business case comes back to a more holistic approach on the total set of solutions and network types that 5G could satisfy.
Can you talk about this dynamic in the carrier market, that 5G will make carriers of us all. How do you see the role of traditional network operators in all of this? Will this give them chance to reinvent themselves and drive new business, or will they remain on the fringes?
Corlis: “It gives carriers a chance to reinvent themselves. From a marketing perspective, they are still very much focused on consumer 5G and the race to beat their competitors. Behind closed doors, they are developing plans on how to capitalise on private LTE and 5G in the enterprise space.
“They know they’re going to struggle to make the case for consumer 5G – on the ground it is too costly and the returns will take too long. They won’t have the same issue in the enterprise space and will be able to demonstrate substantial incremental value to those customers.
“So you will have carriers that offer that service as an enterprise solution, and you will also have some enterprises that want to manage the responsibility themselves, just as they have managed their own wired and wi-fi networks. Private 5G networks are still being defined and will leverage CBRS and other unlicensed spectrum solutions to deliver the desired capabilities.
“In the end, it will be a mix, and operators will have a clear role. Some smaller players will want to have more control and may be more inclined to manage their own networks, while some bigger players may want operators to manage them in order to reduce operational overheads.
“If you have just a single plant, then it won’t be so different from running a localised wi-fi network, whereas a manufacturer with 40 plants across a large geography will find it much harder – and probably require a whole organisation to operate the network.
“So there is justificaton on both sides. The ‘bigs’ might do it themselves, and the ‘smalls’ might too. There’s not enough clarity, yet, to know which way it will go. It’s our job as consultants to help them figure out the most appropriate operating model to meet their business demands.”
Logan: “It won’t be a ‘one-size-fits-all’ model. It will be largely determined on a case-by-case basis, according to the size, capability, and complexity of each organisation. The challenge is similar to any large-scale technological rollout in the manufacturing environment; it will become a question of defining the value-add of moving to 5G and the way to deploy it.”