What do enterprises think of 5G? Frankly, they don’t give a damn  

It is just another networking technology, after all; Appledore Research tells of the trouble with the telco sector’s ‘god complex’
About that headline; the message is enterprises do not care about 5G, just for the sake of it – just because it is 5G. It is not a special case, with a special place in their already-tangled net-working estates. It is just another tool, after all – not even ‘another G’, as the fiercest tel-co-obsessed critics have it. It speaks of the telecom’s sector’s god-complex, as if there is some messianic industrial power inside this new technology. The notion 5G should be gifted by the old dogs of telecoms for indus-trial change is anathema to enterprises, the story goes.

So says market advisory firm Appledore Research, taking a hard line on the task for traditional telecoms to crack the Industry 4.0 market, and get anywhere near the ser-vice-level agreements (SLAs) that will go to underpin it. What do they think of mobile op-erators? Oh, to be a fly on the wall. As it goes, Francis Haysom, principal analyst at the com-pany, retells a story about a faceoff between the sides in an anonymous meeting room, where the telco set, represented by a vendor and an operator, sit mob-handed across the table from a “major” manufacturer.

Haysom – enterprises want Wi-Fi-style 5G networks

“The message to them was, ‘We don’t real-ly like you, Mr Vendor, because we don’t think you understand us. And, as for you, Mr Oper-ator, we don’t even know what you are doing here.” But hold up; let us rewind. Because the operator community is not out of the game, yet, according to Haysom. In another refresh-ingly frank conversation (see also, pages 16-19 and 28-32, and elsewhere), he suggests the best of them – Vodafone and Deutsche Telekom, plus some others perhaps – can carve a role as ‘master builders’, in a few spe-cialist industries in a few regional markets.

But the SLAs they write will not be about “green lights” on industrial 5G componentry; they will be about managing disaster recovery in wider-reaching enterprise sys-tems, which go horizontally across the shop floor and vertically into the cloud. Instead of networking knowhow, the opportunity for mobile operators is to bring their scale to bear on this discipline and push support costs down, as critical workloads are put on-line locally in edge networks. Which means, the logic goes, centralising their support functions, rather than dispersing them into regional enterprise NOCs, or into customers’ premises, as Vodafone hints at (see pages 34-38). That is the gist of the conversation, laid out below, in full. Haysom is joined by Rob-ert Curran, consulting analyst at the firm, in equally forthright mood.

So much with 5G is riding on enterprises, and yet the narrative around ‘vertical’ spectrum and easy 5G edge networks says anyone can be a mobile operator. So where does that leave mobile operators?
Haysom: “You have to look at what enter-prises want, which is for these 5G systems to look and behave like Wi-Fi networks, even if they do all this clever stuff, as well. They want private 5G to be operated by their own network teams on site, which are already running these other factory networks. Because 5G is just another access technology that plugs into the enterprise network.

“Think about what Juniper is doing with [AI networking startup] Mist (acquired in 2019 for $405 million) and AI-driven wireless LAN [local-area network] systems; that is the way it is going – so the whole management headache is removed. Enterprises want the same with cellular. They don’t want a NOC; they want a self-optimizing access technol-ogy that works out-of-the-box, which they can run and manage on their own.

“So it is not about 5G, as such. And for mobile operators, the challenge is whether they can be better (or cheaper) managers of these networks. Like with BT, for example; its global services business will run the NHS network. Not because BT is the best ether-net provider, but because it manages the network at a lower cost than someone else. The opportunity for telcos is not to build a NOC to run these networks, but to use an existing NOC to share the cost of running them – to make it cheaper to run.”

But the NHS has a vast tech estate. Is it the same for a 5G network running in a single factory, or even in a network of factories? Can the carriers scale a managed service down for such a varied customer base?
Haysom: “If it is literally compressed into a box on site, then it is not about the manage-ment of that single box. It is about manag-ing everything else as well – the ethernet and Wi-Fi, and whatever else. It is about guaranteeing quality-of-service across the entire infrastructure, at a lower cost than if the enterprise ran it alone. There is an option, of course, for enterprises to reduce costs further by using the public network, instead of going private for everything – by using the carrier’s edge (MEC) capabilities.

“It all comes down to cost. The big benefit of the public cloud is that it is cheaper – you can pull IT people out of the supermarket, for example. The last thing an operator wants to do is put a person on site, and they don’t want to build an enterprise NOC, ei-ther. Not because a NOC is not needed, but because the enterprise doesn’t want to pay for it. What they want is a cheaper way than having their own staff running the network.”

“It is an open question about where operators will play. But it has to be in that slightly old fashioned systems integration role. Because enterprises see the network as theirs; it is not a special case just because it is cellular”

Francis Haysom, principal analyst, Appledore Research
What is the most compatible business model for private 5G management – to get to a position to write the SLAs against network performance? Is this where the local system integrator enters the scene?
Curran: “Yes. The likes of Cisco, Juniper, Microsoft sell through a vast network of resellers. These national and regional firms have a lot of trust. And these companies are more likely to run some on-site functions for four or five local companies – like schools, or whatever. That is the most compatible model. Because it is localised. The whole premise for (national) mobile operators is their economy of scale, which does not amount to much in this market.

“And even between those, you have T-Sys-tems and IBM, and all of the rest of them – which used to be called ‘system integrators’. I don’t know what they’re called now. But they will do the outsourcing. And then you have Infosys and Tech Mahindra, another level up. So with a SME, you are still multiple layers away from a mobile operator taking on the management of a private network. I just don’t see that happening.”

Okay. So, it feels a bit like we are flogging a dead horse; what is left for them, exactly? Just the Industry 4.0 mega deals?
Haysom: “For mobile operators to be really relevant in this, they need to see themselves as network resellers and not as communi-cation service providers. They have to see themselves as a network management integrator. And therein lies the challenge, like we said. The telco model doesn’t scale down very well. The question is whether telcos will compete with some of their old partners. Because they are used to pro-viding connectivity for IBM to run a global network for Ford, you know? There is going to be more competition, now. IBM can say, ‘Actually, we don’t need your connectivity; we are directly wired together from our data centres around the world. We don’t need AT&T or Verizon, or whoever’.

“So it is an open question about where operators will play. But it has to be in that slightly old fashioned systems integration role. Because enterprises see the network as theirs; it is not a special case just because it is cellular. They don’t want anything to do with the public network. We’ve been on calls where this question has arisen about roam-ing between public and private networks, and there just isn’t much interest. Yes, you can do it, but they want to be separate.”

So with simplified 5G networks, separation from public infrastructure, and familiarity with local specialists, it seems like operators are a couple of steps-removed from hands-on network management. Is it about second and third-line support, for them,
and how quickly they respond in emergency?
Curran: “Yes; it is about business-to-business SLAs, rather than business-to-network SLAs. There will be availability guarantees, and so on. But that is just a portion of the SLA. It is totally different to a telco talking about coverage in its own network, or whatever. In an enterprise context, it is about how well the telco is supporting the enterprise. If the network is working fine, but the enterprise is not getting the changes it needs, then that is an issue – as opposed to just being happy all the lights are green.

“So, you have three data centres, and one goes down – well, you need to plan for that. And you are reliant on connections to Salesforce and Amazon and so on for your internal applications, or for the face you present to customers. That is not about net-work performance; you can’t measure that by looking at green lights on a network. It is about how the organisation is engaged.

“With 5G, there are certain applications that make 5G different – these kinds of high-bandwidth closed-loop visual inspec-tions that use real time data to change how robots move in the factory, or how paint is being used. All of that’s important. But 5G, 6G, 10G? ‘We don’t care.’ Plus, these appli-cations are very niche – you can’t make a whole business around those applications. At the same time, of course, we haven’t done any of this before, because it hasn’t been possible. So maybe it will change.”

“There are certain applications that make 5G different – these high-bandwidth closed-loop visual inspections to change how robots are moving in the factory. All of that’s important. But 5G, 6G, 10G? ‘We don’t care.’ Plus, these applications are very niche – you can’t make a whole business around them.”

How do you view their chances at making a go of this? Will all of this private 5G management be hoovered up by Accenture and IBM?
Curran: “I mean, telcos just completely missed IoT – industry events and confer-ences, where there was lots of innovation, and no mobile operators. It was strange. They are kind of waking up to private net-works, but they are pretty late. They won’t admit it. And it’s only because of the decline in other things, and the flattening-out and dropping-off in revenue. So they have to go somewhere, and here is something that kind of looks familiar, and is worth chasing. Some companies are making a strong play. Vodafone is pretty notable; and probably Deutsche Telekom in Germany, but only in Germany. A few operators will take a lead in a few geographies, and the rest will be out-maneuvered by IBM and Accenture, and local enterprise resellers and integrators.”

So some regional, vertical bright spots, then?
Haysom: “Yes. It is going to be regional. There is an opportunity for them as resellers and managers of industrial 5G networks in countries with strong manufacturing – so in Germany, China, Japan. But they can’t do all sectors; they need to be experts in indus-try, and not just in 5G. So it makes sense Vodafone and Deutsche Telekom will make inroads in Germany, probably. But the telcos that see this as a continuation of what they do in the macro network will fail. They will need, almost, to create different companies to supply industry, which won’t have much to do with their telco businesses – like BT with its sports and television businesses.”

What about Nokia, out of interest? Is it making a fist of things, to be able to go from straight kit supply into managed services? Or will it work via these others, as well?
Haysom: “Nokia is the one that could make the transition. But again, it is not Nokia’s reg-ular managed services business [doing this]; its dedicated private networks business [will offer] managed services or system integra-tion. If you compare with Ericsson, I think Nokia has made the right decision; it has recognised private networks is a different market. That is probably the most important decision the company has made.” EnterpriseIoTInsight