Put a pin in the bombast and a rinse in the dazzle; if MWC is an industry waypoint on the telecoms road, then the 2023 version said private cellular, as a bountiful diversion, is probably just about on the horizon line. Unlike two years ago, when private LTE, in a shrunk-down enterprise format, was only really part of backroom discussions, and last year, when it was usurped on the floor (if not at the table) by private 5G, then this year it was all over the place – still madly hyped, but also reasonably available, somewhat deployed, and variously understood.
The fact the show sought, on stage, to disentangle the hype and reality might even suggest the discussion about private cellular has come full circle – to more urgently consider what it does, where it fits, and how it makes a difference. Such discussion tends to lean on influential non-telco voices, carrying increasing weight as private 5G go-to-market strategies take shape. Indeed, if MWC is a stopping point on the road, then the caravan paused long enough this time for certain outsiders to climb aboard – and maybe even to give it a more serious look.
The most conspicuous new riders at MWC were Hewlett Packard Enterprise (HPE) and Cisco, both from the IT side; we will return to Cisco later (in conversation also with NTT). But we called HPE’s late passage on the 5G train early; its move to acquire core network specialist Athonet, announced just prior to MWC, lights the MWC touchpaper on private 5G, we said at the time. Game, set, and match – and the contest hasn’t even started, we wrote. Our view has not changed between times; the deal, when it closes, gives HPE a first-class ticket – to wherever the private 5G journey leads.
Twelve months before, HPE sounded like it was either unsure or unable to catch a ride; speaking with it now, the message is it was just ill-equipped. The Athonet deal hands it a well-furnished kitbag, and a seat in the dining car. Speaking with RCR Wireless – actually, on the end of the line from Barcelona, as (some of) RCR Wireless was detained at home – Stuart Walker Strickland, chief technology officer for the company’s wireless division, suggests an integrated Aruba-Athonet product will be out next year, with a stop-gap Wi-Fi-like 5G product out sooner.
Indeed, it is the same key message as last year, and as also from Cisco, about Wi-Fi co-existence and integration. But the strategy is in sharp relief; about markets and applications, and channel development. Wide coverage and variable control are key industrial applications, notes Walker Strickland, but HPE, like certain others, is betting big on neutral host opportunities, somewhere down the line. “If it takes off, then 5G moves from being a special case to the preferred solution for indoor wireless… and it becomes a completely different economic game,” he says.
What else? HPE might some day (hypothetically) get into the RAN game, but is happy with what is available from partners today. The Athonet brand will likely (unconfirmed) be subsumed somehow into the HPE family. Most strikingly (or not; “I don’t think it’s really a debate,” says Walker Strickland), he questions the value of mobile operators in the enterprise channel. Looking around at the other riders in this enterprise carriage, he does not see many, if any, operators there. It is an interesting discussion, and worth printing in full; all the responses are from Walker Strickland.
…
How is the show from an HPE point of view? Does it feel like the telco world has moved around to the HPE worldview a little bit – to put sufficient focus on enterprises?
“Yes; well, partially – it’s not going to happen overnight. But private cellular is a big theme of the show. You can see that on the signs everywhere. Which is encouraging to see. There is still, in my view, misplaced faith that the operators are going to deliver private 5G to enterprises. We are convinced that it is going to be the companies that normally serve the enterprise, like ourselves, which will deliver private 5G to the enterprise.
“And that one of the big developments that will make that possible now in a way that wasn’t before is spectrum which is not dependent upon operators for customers to use. CBRS has been a real breakthrough in the US – giving customers with a long-standing interest in cellular, who have never been able to play with it before, a chance to turn it on in their own environments, and figure out what works and what doesn’t.
“The next step for us, now some of them have that experience, is to make cellular available to them in a way that is easily digestible and accessible, and familiar to an enterprise. Both from the point of view of the management of the system and also in a way that works well with client devices that support both cellular and Wi-Fi.”
So just on the telco-view of the enterprise networking space – are there things that are being missed in the way it is understood and presented at MWC, and beyond, and being sold in the market?
“There is an over-optimism, a hard-to-justify optimism, about the role operators will play with enterprises. I know a lot of the 5G project, in the minds of operators, was designed to open this market. And from a technical point of view, they have done a lot in the 3GPP specifications to make 5G attractive to enterprises. But there is still a relatively unsophisticated understanding of enterprises – how they work, what they need, what their existing networks can do.
“Because in many cases, their networking capabilities are much more sophisticated than something like network slicing in 5G. The things a private cellular network brings to an enterprise are not necessarily the things that the operators bring to the table when they approach enterprises. And there may need to be some change in their thinking if they want to be successful in this market.”
I caught up with HPE at MWC last year, and it was talking about developing its own core product. Is that ongoing? Has that work been abandoned?
“Yes, so the announcement last year didn’t have much to do with the Aruba business in the enterprise market. It was more about a longstanding HPE project to develop a cellular core, to be offered in partnership with other players. HPE’s expertise is in authentication servers and user data management, and identity generally. It has relied upon others to complete the rest of the mobility management and all the other stuff that in the… 5G core.
“It has been dependent upon other players to put all the pieces together – not just the radio, but also the software. And for a long time, the Aruba team – [in charge of] enterprise wireless LAN and switching and edge compute – had wanted to integrate with the rest of the HPE cellular business. But the parts it most needed to integrate-in were in the hands of partners and not in-house. The Athonet acquisition [will] change that.
“With Athonet, we get these core components that are essential for comm back-and-forth between cellular and Wi-Fi – so we can make those systems work, and configure and manage in a way that’s very similar to how enterprises are used to managing Wi-Fi. The main difference is that we would [otherwise] have had to work with partners who weren’t properly motivated. Now the plan is for this to become an integral part of HPE’s portfolio.
“And so the announcement last year was more about the intention of the HPE core-network side to focus its attention on scaling products for a private market. But what we’re doing now is really a game changer. We are going to integrate that [core network from Athonet] in with Wi-Fi and deliver it to a private market through the same enterprise channels we’re already familiar with – rather than waiting for the operators to do that.”
Does that mean the project to expand the existing core capability within the business, announced last year, has been sidelined or usurped – not the business, but the idea to expand it?
“They are different and independent objectives. I wouldn’t say the other has been de-emphasized. [For the Aruba team] it continues, but it continues with the Athonet assets – which will come fully under HPE’s control – rather than via partners. The channels are very similar to the Athonet channels – often in licensed spectrum, often with operators, or else on very large sites where licensed spectrum has become available.
“Which means public transport sites or airports, for example, or large defence-related networks. In addition, we want to go after more typical enterprises, including everything from K-12 schools to university campuses, carpeted officers, and industrial sites and factory floors. But those require a different go-to-market strategy than just working through carriers for airports or public transportation systems.”
You mention operators, and there is clearly a debate, which they are also having with themselves, about their role in all of this; but HPE, like Athonet, has important partnerships with them, and would presumably rate them in certain channels and applications. Just set out how you see their role in the supply of private 5G to enterprises?
“So, yes, there is a very constructive partnership to be had. One of the most compelling use cases in the long term is going to be providing neutral host coverage within enterprises, especially indoor enterprises – which is, essentially, roaming from the public network onto the private cellular network, which is helpful and necessary because we’ve got a lot of customer sites where there’s very poor cellular coverage, but no appetite for DAS system, and no real movement from operators to solve that problem in any other way.
“And so our customers are prepared to take that into their own hands by deploying a private cellular network. But for that use case to work, we need agreements with operators to support that inbound roaming – and outbound roaming as well, for that matter. We have had a project in place now for a couple of years – Aruba Air Pass, based on Passpoint technology – which has been very successful on a trial basis in North America. It supports all the major operators, but only in about 20 or 30 customers across maybe a couple thousand sites.
“So we know that there is an interest, even among the operators, to support that idea – to extend their footprints through infrastructure that has been deployed by private companies. But it has been difficult to scale globally because it relies upon Wi-Fi, and is based on Passpoint technology. And different operators in different parts of the world feel differently about Wi-Fi, and Passpoint isn’t widely supported.
“So our hypothesis is that by using a private cellular system based on technology that operators are familiar with – whose interfaces are native to what they’re already doing, roaming between each other – then they can roam onto the HPE enterprise footprint as easily as Vodafone customers roam onto AT&T in the US, say. That is a very constructive relationship for them to have with us in this environment. It means additional-footprint for them and it solves a coverage problem for our customers.
“Where I am more sceptical about the role of operators – and I don’t think it’s really a debate – is their ability to be a channel to the enterprise for private 5G. My only argument about that is that they’ve had an opportunity to do that for a long time, and they have made very little progress outside of a few specific industrial vertical markets. And I think that will continue. Whereas we have really strong relationships with enterprises, and others also have a really strong relationship with them. It has only been hard in the past because we relied upon operator spectrum.
“Maybe someday there will be an efficient way for operators to make their licensed spectrum available for private usage, but that’s not the case today, and it’s a cumbersome and difficult process which usually ends not-happily for anyone. This [vertical private] spectrum in a growing number of countries gives hope that enterprises will be able to have a truly private network under their own control – independent of operators, but using cellular technology. Because they are interested in the technology, and not in complicated relationships with operators.”
Can you clarify HPE’s position on the RAN piece? Is that something HPE will always partner on, or is it something it is looking to develop also?
“So when we started looking at this, we thought we could add most value in the integration of management systems and the network, and getting visibility of quality-of-service across the two; so mobility management and identity management identity, for example – bridging the disconnect between cellular systems based on SIM identity and corporate networks based on roles. So figuring out how to sort all of that out. Which is software stuff above the radio.
“By contrast, there is less opportunity for us to differentiate with the RAN. So as with previous offerings from HPE and from Athonet, we will continue to use partners for the radio – and I say the radio, but there is a lot of software in the eNodeB/gNodeB, as well. But over time, I think we will look seriously at whether or not we wish to build our own radio. But that’s not the problem we have to solve in order to address the market.
“There are good options today in that space, which we are comfortable using. We will maybe revisit at a later stage, to see whether or not it makes sense. But it is a complicated thing because there are lots of different regulatory environments. Radios need to be qualified and certified in lots of different places. It’s a more complex world than Wi-Fi, where we basically have a relatively small number of products to cover most of the world.”
Which markets are you most focused on? All of them, or are there either certain more 5G-suitable or HPE-specific segments you are going after?
“There are a couple of different things we’ve noticed with [enterprise cellular in] CBRS over the last couple of years. One thing is [that cellular works for] large coverage areas, typically outdoors, but not always; sometimes in large warehouses – where it would not be cost effective to lay cable for Wi-Fi nodes. And although each individual cellular small cell is more expensive, you can more cost effectively cover a wider area that way.
“That’s one thing – which is partly about cost, but also partly related to quality-of-service issues. Some customers have latency sensitive applications running on mobile devices that travel at speeds faster than a pedestrian – a robot in a warehouse, or a vehicle, or a rollercoaster at an amusement park where it’s critical to get data continuously to that device, and where handover from one Wi-Fi access point to another on a best-effort basis isn’t very satisfying.
“Cellular was designed for high speed handovers – and probably wouldn’t require any handover anyway because of the wide-area coverage. Another vertical focus is where customers want to segregate sensitive data – sometimes for security reasons, but more often for deterministic quality-of-service. Which is the case in an airport, say, where public Wi-Fi traffic ebbs and flows, and you want another network for downloading telemetry data from the aircraft at the gate. The third area – which is more of a promise than a reality so far – is for neutral host infrastructure.
“The offer of wide-area coverage is important for lots of our customers – for university campuses, for example, and many industrial and warehousing sites. It’s really important. The ability to segregate traffic is a more niche industrial use case – mostly in privately managed networks in publicly accessible spaces. Casinos are another example – there’s a backup house network and a public network. But I still think those sorts of cellular use cases represent a relatively small proportion of our overall enterprise networking business.
“But the neutral host model is relevant, potentially, to the vast majority of our enterprise customers. If it takes off, then 5G moves from being a special case to the preferred solution for indoor wireless coverage in most of our customer sites where there’s poor coverage today. And then, you know, the volumes for cellular begin to approach the volumes for Wi-Fi, and it becomes a completely different economic game. If we start to move from public cell tower volumes to almost consumer electronics volumes, then the economies of scale start to kick-in in some really interesting ways.”
And what about the more hard-nosed Industry 4.0 venues, on manufacturing production lines – are those on HPE’s radar in the short term?
“For sure. But a remarkable number of these industrial applications are being well-served by Wi-Fi today. A lot of large-factory customers are running their shop floors on Wi-Fi without any cellular technology. So I wouldn’t want to discount Wi-Fi for a large number of those venues. But, yes, manufacturing is also a frequently-referenced use case for private cellular, and we have a lot of interest from that segment – to see what they can get out of 5G.
“I think, generally, there’s been a move recently to wireless – as opposed to wired – control of robotics and things like that in areas where we didn’t think wireless, cellular or Wi-Fi, would work at all. And we have managed to [Wi-Fi] get latencies down to really usable levels. It will depend partly on timing. Some use cases will continue with Wi-Fi because they know the system; newer ones will start with a 5G solution, as the first time they use wireless.”
Is there pent-up demand for private cellular from your existing base? Does that, alone, provide a sufficient uplift in terms of private 5G sales, now you have this additional capability? Or is cellular very much a driver of new business for HPE?
“We have a substantial market share in enterprise for our existing network products. So it will certainly be a priority to make sure those customers get what we have to offer in the 5G space as well, and get it in an integrated way. But it could also be a driver of new business; it is quite possible that 5G will bring us to new customers – which are currently purchasing from competitors who don’t have a 5G offering, and don’t have one that’s complimentary.
“Whether that means getting a foot in the door with the 5G system while there’s a competitive Wi-Fi and fixed network system in place, and then eventually thinking to replace some of it, or whether we approach the market once we have the fully integrated solution – it is too early to tell.”
Will HPE go via system integrators, which are also making a noise about the supply of private cellular systems, or will it go directly to enterprises itself? Just clarify your go-to-market strategy.
“System integrators are a really important part of our channel; we work with a lot of them on Wi-Fi, and they have a great deal of expertise with radio planning and deployment. We rely upon that ecosystem a lot. There’s less of that [channel setup] for cellular, so that’s another thing we’ll develop. But they are not typically the channel to market for us, as such; they provide support in the deployment. Our approach is to be very close to the customer, ourselves.
“So we’re not so much thinking about a model in which a system integrator has a relationship with the customer and they pick the pieces and stitch it all together and deliver it and manage it. We’re thinking of a model in which we have the relationship with the customer; we figure out what they need, get it all together, and make it available. And then a system integrator may well be involved in the deployment or even in the management after that.”
When will a fully integrated system from HPE be available?
“There would be an appetite as soon as we could make it available. But realistically, we’re not talking about a fully integrated product in this calendar year. I’m hopeful we’ll have some of those systems nxt year – at least operating on a trial basis with a number of customers. But it will be quick and straightforward to offer a standalone cellular system that has the look and feel of a Wi-Fi system, even though it’s separate, as an intermediate stage in between.”
Will the Athonet brand be retained?
“I don’t know. But you will have noticed that there’s been a consolidation of brands within HPE under the HPE name and logo, and we are all one company and the plan is to operate as one company.”
Is HPE quite relaxed about its timing? Could it have been quicker? Some have been pushing private cellular hard for some time; some industrial tech companies say private 5G is a 2024/25 market, and not to rush. What is your view?
“I wish we could have done this earlier. I’ve been revisiting the prospects off and on, on a recurring basis, for at least eight years. And there were points earlier when it was just not the right time to do it. I think all of the pieces are in place, and at this point, anyone who is thinking maybe they should wait to see how it pans out in 2025, is going to miss the mark.”
Were there many others in contention?
“We looked at a wide field. There are fewer players in this space today than there were a few years ago. And we looked at all the ones that we thought were suitable for integration into an enterprise-appropriate offering. We have a [resale] relationship with Celona, which we’ve had for a couple of years – which has a nice proposition, also scaled for enterprise, but which is a standalone network; it is convenient to deploy, but it is not integrated. I really felt the Athonet capability and vision stood out as most closely aligned with where we wanted to go. We’re not going to need to twist its arm to convince it this is the way to serve the enterprise market. We have the same vision.”
