A couple of interesting-sounding mission-critical slicing exercises by mobile operators in Europe in the last week or so got RCR Wireless to thinking about whether the slow rollout of standalone 5G (5G SA) in national network infrastructure has reached a point where it is about to give this once-hyped technology a real lease of life. Does new work by Hrvatski Telekom and Drei Austria – to hive off airtime channels in their national 5G networks for port operations and emergency services, respectively – somehow suggest that network slicing is ‘coming of age’, finally? You’re having a laugh, right?
That was the response from the analyst fraternity, quicker to respond to the outreach that followed from the question than the operator community – whose responses may yet form a separate piece. The question was over-egged, clearly, just as the notion of network slicing was when 5G appeared five years ago, attached to 4G-LTE bearer networks in non-standalone (NSA) mode – and as it has more or less existed ever since. But as Ericsson says, it is only “half time in the 5G game”, and sliceable 5G SA is starting to replace derivative NSA systems. So has something changed?
Will mobile operators now start to bring the full force of 5G SA technology in public network infrastructure to bear on the enterprise market? Are the Croatian and Austrian projects examples of how they can reassert the value of their national airtime assets in diverse enterprise markets – having been practically discounted from the running as private cellular in private spectrum has filled the hype vacuum long-since vacated by public network slicing? “That seems a little premature to conclude,” responds Robert Curran, consulting analyst at Appledore Research.
True, true; but something is happening. Let us recap. Firstly, port operators APM Terminals and ENNA Group have tapped Deutsche Telekom subsidiary Hrvatski Telekom for a ‘private’ slice of its public 5G network at their joint-venture development at the new Port of Rijeka, set to be the largest shipping port in Croatia when it opens in 2025. Hrvatski Telekom will deploy new radio (RAN) infrastructure at the shipping terminal, but the service will be managed on a ‘slice’ via its public core 5G network. There is no word on spectrum frequencies, as yet.
Secondly, critical comms provider Frequentis has taken a slice of Drei Austria’s national public 5G network in its home country to develop and test applications for public safety and emergency response control centres – as used by civil and military air traffic control, air defence, emergency services, port authorities, and railways. The setup has also seen Drei Austria, in combination with parent CK Hutchison’s global enterprise services division Three Group Solutions, deploy a dedicated private core network at Frequentis’ headquarters in Vienna.
On one hand, the first appears like a rather parochial test project in a greenfield port, and the second is like an updated version of familiar emergency services setups in bigger markets. On the other hand, they both play with slices of public 5G networks, and, importantly, they are both geared towards the kinds of mission-critical (mostly) Industry 4.0 services that tend to prefer all-edge private 5G systems. There are other recent examples of network slicing – for music festivals, boat races, broadcast services – but they don’t rub-up against hard-nosed Industry 4.0 in the same way.
So, what do we think? Appledore Research is quick and expansive with its responses. Has the value of public 5G for enterprises been somehow under-played, recently? “No,” retorts Francis Haysom, principal analyst at the firm, joining Curran in response. “No enterprise has a 5G problem. No enterprise has a network slicing problem. They have problems that 5G technology might support. Until telcos understand this… it will never come of age. The reason for the underplay is [a response to] the overplay – about 5G being a solution that enterprises want.”
Dean Bubley, founder of Disruptive Analysis, picks apart the scant detail about the Port of Rijeka setup. “It looks like a greenfield port project, so presumably they can design the whole wireless environment from scratch as well. Slicing is potentially viable in situations where there is a lot of dedicated and custom RAN engineering for a particular site – so sufficient coverage and capacity can be proven in advance, including for applications inside warehouses or across waterfronts. It is not really viable from the ‘normal’ default macro RAN coverage,” he says.
But there are so many questions, anyway; about the spectrum APM Terminals is using, whether its port activities are to contend with “non-port RAN nearby”, whether trucks and lorries will use the slice(s) across the rest of the country. Which is arguably the primary use case for slicing a public network – to support nationwide mobility applications. Curran says: “We expect [public 5G slicing] to find niches – like in transportation, logistics, emergency services, where there might be some advantage in tapping into nationally available infrastructure. But plenty of firms are happy on 4G.”
His colleague, Haysom, is to-the-point. So why is this happening now – if, indeed, it is? “Because vendors need to sell the rationale for moving to 5G SA,” he responds. What is the timeline for slicing to be a real commercial and strategic proposition for enterprises? “There is no timeline without [operators] changing to an enterprise viewpoint,” he says. What’s the likely commercial model? “Nobody will buy a slice – they will buy a solution to a problem. Commercial models will reflect the nature and the cost of the problems they solve.”
But the original question, actually, about whether sliced public 5G SA will have its day, is too narrow. It needs to be divided into two parts: about public 5G SA for enterprises, and separately about the slicing of it. Europe and America are not necessarily bellwethers for the value of public 5G SA technologies to enterprises, says Dave Bolan, research director at Dell’Oro Group; their progress with this shows only how few 5G SA networks have been deployed in these regions, and, as per the Appledore narrative, how unreconstructed their mobile operators are.
“The value of 5G SA networks for enterprises has been well-established in China since 2020 [where] the market has chosen to use MNO-provided MPN architecture,” he says. As a note: MPN is short-hand for (mobile) private networks; but so is PMN, to swap the order, and so is NPN, in 3GPP parlance, which means non-public networks and corrects the idea to use just PN for private networks. There are other acronyms, too (DPN, P5G). In his explanation, Bolan combines MPN with MNO and MEC, in the context of ROI – which don’t need regurgitating, but make for a mouthful.
His point is that China has a good record already with MNO-MPNs using 5G SA connected to public MEC points – to deliver ROI. “China has been highly successful [proving] ROI over various applications and industries in one-to-three years. Its 5G SA networks are very mature, with public MEC nodes sprinkled liberally around the country, providing low latency… for broad geographic coverage. MNOs are also geared up to implement MNO-provided MPN on-prem apps quickly. The market has developed an ecosystem of solutions for enterprises that can be quickly replicated.
“Unfortunately, this has not happened outside of China for several reasons: slow take-up of 5G SA by MNOs, with most defaulting to 5G NSA; lack of high-performance ecosystem solutions that need 5G SA; and [because] ROI needs to meet market requirements – two-to-three years is too long for many enterprises. To date, 28 countries have 5G SA networks, and most attention has been paid to the consumer rather than enterprise market. As 5G SA networks mature, attention will move toward enterprise applications.” So outside of China, it’s just about (im)maturity, then.
But the second part, about 5G SA slicing, prompts a more circumspect response – which echoes the other ‘slice-deniers’ and carrier-agitators in this article. Network slicing is not required for on-prem Industry 4.0 applications, anyway, says Bolan. So where’s the value? Mobile operators can ensure low latency and high reliability already, he says, by supplying dedicated on-prem radio (RAN) and edge (MEC) equipment – even if they attach to public radio spectrum and a public core network. “Only the designated enterprise has access to the RAN and MEC,” he says.
He echoes the point about public network slices for national mobility applications. “Network slicing is needed on a shared RAN and public MEC node for enterprises that require broad geographic coverage – such as for logistic companies and public safety.” Which almost, maybe, describes the Hrvatski Telekom and Drei Austria exercises. But the problem for mobile operators – everywhere, including in China – is that the developer ecosystem has to be harnessed in order to produce enterprise solutions to attach to public 5G networks, or slices thereof.
“The MNO community has realised that more has to be done to accelerate the uptake of applications, and that means attracting the application developer community to make it worth their time to invest in applications that can be leveraged across many MNOs networks and the enterprise market,” says Bolan, pointing to the GSMA’s Open Gateway initiative to provide a framework of common network APIs to open general access to operator networks, and subsequent work with the open-source CAMARA Project to define, develop, and test the APIs.
He also mentions the role of reduced capability 5G (RedCap) just to get prices of 5G SA hardware down for IoT projects. “It will take several years for all of these initiatives to mature and be leveraged across many MNOs, enterprises, and applications. Until then, we will continue to see proof-of-concepts, trials, and one-off demonstrations of what can be done,” he says. In other words: no, not yet; there’s no coming of age. Are you having a laugh?
