Spectrum, systems, slicing – five key challenges to bring private 5G to the masses
A session at Private Networks European Forum (PNEF) last month considered the challenges, and broader cultural shift, to bring ‘private 5G to the masses’. Koen Mioulet, secretary at the European Users Wireless Enterprise Network Association (EUWENA), and Maria Cuevas, networks research director at UK-based telecoms operator BT, took questions about regulatory and technical issues, respectively.
Together, they also addressed how the market should rationalise excitable ‘vertical’ workstreams, geared to define siloed 5G use-case solutions for different industrial sectors, to come up with a master ‘blueprint’ for Industry 4.0, and how it might do better to collaborate and consult with the broad industrial market, including small and mid-sized firms (SMEs) as well, to raise awareness and ensure private 5G is not just a plaything for rich mega-corporations.
This was the gist of the session-talk at the PNEF event, hosted by RCR Wireless and Enterprise IoT Insights; it has been transcribed and reordered, here, in list-form to provide a take-away summary of the state of (industrial) ‘things’, as connected and animated by new private LTE and 5G networks, and how this novel network tech might just spur a ‘fourth industrial revolution’ – as the market likes to imagine.
So, here goes; five big challenges to bring private 5G to the masses, as defined at PNEF by Mioulet and Cuevas, and their distinctive, but complementary, lobbying and research work at EUWENA and BT.
1 | SPECTRUM HARMONISATION
The first challenge is spectrum harmonisation, to make so-called ‘vertical’ bands for private 5G uniform across markets. This was made light of in the opening remarks at PNEF, on the grounds spectrum access has been cleared in many markets, and national enterprises have the same work to navigate local regulation, however it is matched elsewhere, and multinational enterprises have the regulatory wherewithal to go market-by-market, without bother.
But this also misses the point; that fragmentation of spectrum cascades down into the telecoms supply-chain, and makes the vendor-mix and architectural-match, around radio-gear and terminal-devices, more complex for enterprises. EUWENA’s whole raison d’etre is spectrum harmonisation for enterprises; it wants consistent regulation across Europe, and ideally across the world. Mioulet, secretary at EUWENA, made clear at PNEF that
He said: “Spectrum is still organised per-country. Let’s take our best example: CBRS in the US – which is a [single] harmonised spectrum-regime across all 50 states. What you do in Delaware, you do in California. Brilliant. In Europe, what you do in the UK is different to what you do in France. There are countries that don’t even have private spectrum, so you’re at the mercy of an operator… [But] if you want to go truly private, you need private-spectrum.”
He noted the European Union (EU) is pushing to follow the UK example, to liberate 400 MHz spectrum for dedicated enterprise 5G usage at 3.8-4.2 GHz; by contrast, Germany, most quickly, legislated to make available 100 MHz at 3.7-3.8 GHz, while French Industry 4.0 has until recently been kicking about at 2.6 GHz. But France has just released the 3.8-4.0 GHz band, and the EU agenda is hopeful. “It will improve over time. There is no doubt.”
But the point stands, as well, that regulatory chaos reigns in Europe, currently. “Holland is even worse; we haven’t yet come to a decision,” said Mioulet. “In all fairness, spectrum is a definite issue.”
2 | SPECTRUM LIBERALISATION
The bigger issue, however, is just with spectrum liberalisation – whether as a local-market precedent, to break old monopolies to liberate airwaves for enterprise usage, or as an expanding economic policy, to go beyond parochial experiments with slivers of mid-band spectrum in the name of ‘innovation’. It applies to the Netherlands, and other laggard regulatory markets, as Mioulet mentions, but also to markets with private 5G provisions already in place.
The first point is that mass-market private 5G is a non-starter, anyway, without spectrum. Referencing a previous PNEF address from GSA about the run-rate for private 5G deployments, Mioulet said: “Germany and the UK are doing well… [because] the second you liberate spectrum, the market follows. In Holland, the market is hampered by the fact that spectrum hasn’t been allocated yet – [the same in] Spain, which hasn’t allocated spectrum either.”
He added: “Over time, things will harmonise, of course. But… the GSA presented 889 [deployments], which is impressive on one hand [but] you would imagine that it should have been thousands, by now. So Germany has over 200 [deployments], already, [and] so does the UK (actually the GSA puts Germany and the UK at 77 and 39, only) – well, every other country could have 200-plus systems if only the spectrum were available.”
The second point is that, even where airwaves have been pegged for local enterprise usage, more spectrum will be needed, eventually – if local installations are to be loaded with Industry 4.0 apps, and private 5G is to go bananas. Which is why New Telco and Old Telco are at loggerheads over CBRS in the US, and the idea to extend the model (by 350 MHz, to 500 MHz in total) to the 3.1-3.45 GHz band, to both monetise telecoms and digitalise industry.
Mioulet explained: “A lot of projects are still pilots, consuming maybe 20 MHz or 40 MHz. They are not using 100 MHz or 500 MHz, yet. [But] EUWENA surveyed a couple of tenants at the Port of Rotterdam – crane operators and terminal operators, also refineries and industries – [about] their future demand for wireless, and extrapolated that to over 3,000 tenants on the port… and [it] translates to 500 MHz… Which is more than most countries are allocating.”
He continued: “Of course, [the solution] will aggregate private spectrum, public spectrum, unlicensed [MulteFire] spectrum…. But it was an exercise just to get a feel for the kind of a demand we might see in a large industrial area like a port… [and] what countries are currently allocating [is not enough].” The proposed 400 MHz tranche at 3.8-4.2 GHz in the EU, already allocated in the UK, starts to “look like a decent and useful chunk”, he said.
3 | ECOSYSTEM ALIGNMENT
Another related aspect to spectrum harmonisation, very quickly, is with fragmentation in the ecosystem, and how to knit-together a coherent ‘blueprint’ for private 5G that goes across industrial sectors. This is EUWENA’s parallel mission; to knock-heads in the supply chain so development work is not duplicated, effectively – conducted in private silos for niche ‘vertical’ ends.
“There are all these individual associations, per country [and per vertical],” said Mioulet, making specific reference to the disjointedness of 5G work-streams within the Spectrum Policy Forum in the UK and the 5G Alliance for Connected Industries and Automation (5G-ACIA) in Germany (plus something-unheard in France), and of parallel interests at the European Utilities Telecom Council (EUTC) and The Critical Communications Association (TCCA).
The 5G Automotive Association (5GAA) might be added, too, presumably, as might any number of others – such as the Small Cells Forum (SCF), the MulteFire Association (MFA), the Industry IoT Consortium (IIC), and the Wireless Broadband Alliance (WBA), all issuing ‘blueprints’ on private 5G in one form or another; alongside private vendor adventures up the same garden path by the likes of Nokia and Qualcomm, most notably.
“We need to harmonise the ecosystem, as well. We need to get away from all this fragmentation; everybody just looks at one vertical or one country, [and it requires] closer aggregation across both,” said Mioulet. It is bog-standard industrial-grade self-absorption – a classic tale of “the perceived uniqueness of one’s own vertical”, he reckons. “Right? Mining is inclined to think that mining is unique; ports are inclined to think that ports are unique.”
“Because nothing resembles a port. But if you take a helicopter view, there are always similarities… No sector is that terribly unique. Just anecdotally, Amsterdam Airport (Schiphol) and the Port of Rotterdam engaged a couple of years ago – from a start point where they both said, ‘Oh, aren’t we unique?’ – and guess what; after a couple of meetings, they found they are both covering large areas, they both have critical KPIs, they are both moving metal objects.
“And in the end, they said, ‘Hey, our cases are somewhat similar’. [There is a need to look] at similarities across silos, to get away from this perceived uniqueness. That is one of our concerns, you still get 5G-ACIA looking at manufacturing and EUTC looking just at utilities… So yeah, it is still very siloed. We need to get away from that, because otherwise the market will always stay fragmented.”
4 | TESTBEDS / DEVELOPMENT
Over to BT; Maria Cuevas, the company’s networks research director, driving at the same thing, effectively. Cuevas told PNEF that cross-sector collaboration is not just required to raise private 5G projects out of trial, but to extend trials as well, to drive integration and innovation. Indeed, it is worth celebrating the development work that goes in at the start, she said – and not getting hung-up about premature scaling of private 5G systems for the mass market.
It was a riposte, of sorts, perhaps, to the PNEF meta-narrative, discussed in the intro/outro sessions, and at the end of the GSA presentation, about whether Germany’s private-5G run-rate should be asterisked – disqualified or contextualised – on the grounds it is mostly obsessed with hard-nosed Industry 4.0, and therefore practically stalled, in large-scale experimentation, pending future releases of the 5G standard. Maybe not; but her argument fits.
Cuevas said: “If you look at the pace of the evolution of any other service out there on the internet, that continuous circle and continuous development does exist. So I don’t see why we are over-critical in the telecom industry about when exactly we move from trial to commercial product. But [experimentation and collaboration] is incredibly important in this ecosystem, because private networks are not just connectivity pipes [going] from A to B.
“These are end-to-end solutions, to serve customer needs. Which means, by definition, you need a number of parties involved. It is an ecosystem play; it is not a single-player solution – with integration challenges to overcome. But more important: customers need to understand the value these solutions bring, and be able to play with them in order to explore the art-of-the-possible. Which is what BT is doing. It’s a learning process for everybody.”
This work – much of it at Adastral Park, BT’s research headquarters in the UK, where it runs public testbeds and showcases, as well as private pursuits and collaborations – sounds like another siloed research exercise, which EUWENA might prefer to see in an all-industry mashup. And it is, to an extent – except co-creation is a north star for BT’s own Industry 4.0 travels, and most of the work is about the gnarly tech-workings of private 5G (see below).
Which, as carried-led exercises, developing into industry-wide practices, will bring private 5G to the masses, also. Responding to a question at the PNEF session about whether its trials are to find ways to monetise 10 milliseconds of latency in an all-edge private network, for example, Cuevas responded that they are more about just getting stuff to work properly – as required, as promised, as delivers value.
She said: “At a technical level, we do an awful lot of testing and validation, and trying to understand the needs of different use cases. I don’t think our business is to sell 10 milliseconds. Our business is to understand the needs of all those different applications and obviously offer the best possible solution underneath it. And what you need is the right tools in your box to be able to deliver what’s best for each use case.”
It is a more nuanced point, perhaps – carrier-led technical research, in the face of a discussion (above) about cross-sector rationalisation – but it argues for the same, actually, and forms the root for commercial solutions to flower. As such, it goes in-step with, and even proceeds, everything else here – spectrum access, spectrum harmonisation, ecosystem harmonisation – and leads, for example, to essential system integration (as below), so IoT crosses easily between private and public 5G networks in the mass market.
5 | PUBLIC INTEGRATION
BT is working, specifically, to understand how the public network plays and integrates with private network systems – which despite contrarian arguments, as with network slicing, that Industry 4.0 wants all-edge solutions, will surely popularise 5G in the broader enterprise market. Here we deal with both concepts: the awkwardly named challenge of public-network-integrated non-public networks (PNI-NPN), plus (PN and NPN) slicing.
Cuevas said: “A lot of trials and services will be based on a standalone [private 5G] solution – a system dedicated to a customer, deployed on their site, [in private] spectrum… That will probably be a very good solution for a very wide range of customers. But there are lots of possibilities to expand on that with 5G – which is where the whole PNI-NPN concept kicks-in. [And] we own a macro network [that] we want to leverage to serve customers in the best way.”
She went on: “A logistics solution, say, to track assets on a campus might also continue between sites – and transition from a standalone private network onto the wide-area [public] network. And for that, you want the same level of service. The customer does not care how these services are delivered; all they care about is that the solution works, in their campus, and between their sites as well. So that’s effectively what we are looking at doing next.”
It is a vision, surely, of mass-market enterprise 5G. The elephant in the room, of course, is network slicing – around which there was elephant-sized hype a couple of years ago, but which has skulked in the shadows more recently. So, why should enterprises be excited about slicing? Which is how the question was put to Cuevas at PNEF. Why is it important? What are the challenges? When will we see PN slicing, in earnest?
Cuevas responded: “So, yes, it is still incredibly important, and, yes, it is very challenging – which is why these things take time to mature, and get deployed for mass-market rollout. But, again, the industry is also over-critical, and shoots itself in the foot… [about] why-is-network-slicing-not-out-there. Actually, there are many forms of network slicing already. I mean, we offer emergency service [networks] in the UK… by slicing parts of our network.
“Which means we can provide the right availability, resilience, quality-of-service. You might argue this is not full end-to-end slicing, but… that is a mission-critical service, as we all understand it, and I would probably refer to it as the first proper slicing in our network. With regards to private 5G, slicing will play a role in standalone solutions, to [manage and prioritise] different use cases in large sites – whether it’s end-to-end, or just some aspects of the radio.
“In the PNI-NPN approach – transitioning from all these [NPN] islands, to provide a seamless experience across an entire country – then we are going to need a form of slicing across the public network to be able to deliver a private-network experience to customers when they are outside their campus. Which, again, is where slicing becomes very relevant – in how to allocate radio resources, and route traffic across our networks.”
She goes on: “That might require some form of local breakout, and perhaps some traffic routed to an edge compute node – which might actually be in the customer premise, even if some of the rest of the network components happen to be shared with the rest of your network. [But] network slicing is not a flick-of-a-switch solution, where you just turn it on one day. It’s an evolution of components and the flexibility that allows you to mix-and-match the capabilities.”
Which ties back into that R&D work in Adastral Park – or in whatever carrier/vendor silo, shared and multiplied across workgroups in ever-closer horizontal and vertical industry associations.