5G in Europe: The Commission is living in VR

Introduction

The European Commission’s white paper on network infrastructure makes many comments about 5G networks, and expresses disappointment with the rate of both deployment and uptake. It leans on outputs from the EU 5G Observatory[1] and the 2023 update on progress towards the Digital Decade targets[2]. It suggests reasons why 5G has so far failed to deliver against expectations and proposes possible ways to improve the situation.

However, the paper primarily blames 5G’s woes on perceived failings of the Single Market, rather than acknowledging that the EC itself ignored the complexities and timelines of deployment, as well as drivers of demand. It calls for consolidation of MNOs (mobile network operators) and centralisation of EU spectrum policy, rather than asking itself hard questions about its own technical na?veté and poorly-framed targets. It is living in its own virtual reality.

In my view consolidation may help, but is largely irrelevant to the underlying issues, while spectrum policy harmonisation is largely a very bad idea.

The Commission avoids any self-criticism on areas such as

·?????? Its ready acceptance of industry hype

·?????? Low awareness of the phasing of 5G, with support of key features only in year 5+

·?????? Poor grasp of the challenges in providing 5G coverage in remote areas and indoors

·?????? No analysis of the realistic patterns of demand for immature / patchy 5G

·?????? Lack of focus on enterprises’ desires to deploy their own private networks

·?????? Irrelevant, poorly-defined or easily-gamed KPIs and targets for the Digital Decade

In essence, the European Commission failed to understand what 5G actually is, when different phases would be available, how and where it would be deployed, who wants to use it and from which companies they wish to buy it. It is now ignoring lessons it could learn from its mistakes. This should raise alarm bells about the quality of proposals in the white paper (eg on spectrum), as well as broader 5G / 6G development initiatives.

Lack of understanding of 5G supply, demand and tech basics

A central problem is that the Commission failed to realise early demand for 5G services needed to map to the realities of early supply.

Much of the hyperbole around 2018-2020 stemmed from an assumption that 5G’s most-advanced features would be available from Day 1, along with near-ubiquity of coverage.

It wrongly assumed the industry would supply gigabit-speed, ultra-low latency connectivity wherever users and their devices required it, from 2020 or shortly thereafter. Furthermore, it over-focused on irrelevant drivers of demand that either do not need 5G (such as virtual reality), did not exist (such as remote surgery), or which would take years to design, test and deploy (such as industrial automation systems).

Instead, it should have looked at catalysing demand for more realistic applications that could use early versions of 5G, deployed progressively and patchily, which still worked OK on 4G, and with a bias towards outdoor use. For instance:

·?????? Fixed wireless access (FWA) broadband

·?????? “More of the same” – lower price, higher speed & better capacity for smartphones

·?????? Enterprise private networks, especially outdoor settings such as ports

·?????? Vehicle connectivity, such as for buses, delivery vehicles and taxis / ridesharing

The paper’s poor grasp of wireless technology is well illustrated by this line on page 11: “EU operators are focussing on reusing existing sites for low and mid-band deployments. However, for future upgrades, e.g. 6G or WiFi 6 the required network densification is likely to increase by a factor of 2-3 by the end of the decade”

Given that 6G will likely not even be available until 2030 – and is yet to be defined or standardised – while Wi-Fi 6 is almost entirely used indoors in homes and businesses and not on mobile sites, the comment makes no sense at all. Yes, some densification may well happen, but not at those levels, nor driven by 6G or Wi-Fi.

The paper repeats the common strawman argument about self-driving vehicles as a major use-case for 5G. “Connected and Automated Mobility is expected to be one of the main drivers to decarbonise the transport sector and 5G is expected to be one of its main enablers”. This notion has been roundly dismissed as fanciful – linking 5G to autonomous vehicles might have been common in 2017, but now results in more rolled eyeballs than rolling wheels.

?

Referencing the 5G Observatory’s International Scoreboard, the white paper notes that “South Korea has deployed more than five times the number of 5G base stations per 100,000 inhabitants than the EU, and China almost triple”. Yet it does not acknowledge both countries’ deep deployment of fibre for backhaul, nor their enthusiastic embrace of indoor wireless as an extension to outdoor. China, for instance, has 100MHz of dedicated 5G spectrum for shared indoor use, and over 1 million of its 5G base stations specifically used for in-building coverage[3].

Flawed metrics, flawed solutions

The Commission’s own Implementing Decision[4] on KPIs to measure progress towards the Digital Decade targets, issued in June 2023, defined “5G coverage, measured as the percentage of populated areas covered by at least one 5G network regardless of the spectrum band used”.

That can essentially be translated on “areas where at least some people see a 5G logo on their phone”. It does not imply being able to do anything useful with that 5G coverage, nor for it to be any better than 4G. Neither does it imply that 5G is useful for any applications needed outside populated areas, such as at industrial sites, agricultural areas or along road/rail routes. Even in populated areas, most people spend 80%+ of their time indoors, yet that isn’t mentioned either.

Instead, the paper focuses on fragmentation – either in terms of numbers of operators across the EU, or the differences in spectrum availability and rules.

These are not bad things to discuss in principle, but they are secondary problems.

it is arguably true that Europe could benefit from fewer mainstream MNOs, as some clearly struggle to make returns. It also benefits from having a more diverse landscape of specialists or delayered players. Private networks, neutral hosts, regional fixed-wireless providers, tower companies, industry-specific networks, MVNOs and many other categories can all deploy and run 5G infrastructure, especially as networks become more software-based and easy to build.

In many ways, the first industry transformed by 5G is the telecoms industry itself. In future we should have fewer identikit generic MNOs as consolidation shrinks their numbers, but more specialists and shared infrastructure providers.

Better late than never

Some aspects mentioned in the white paper are showing signs of improvement. It recognises the difference between early 5G (called non-standalone, or NSA) and “proper 5G” (standalone, or SA) which uses a dedicated core network than can support advanced features.

“Most often, where 5G is deployed, it is not “stand-alone”, i.e. with a core network separate from previous generations. Prospects for deployment of 5G stand-alone networks ensuring high reliability and low latency, which are key enablers for industrial use cases, are not good. The deployment of such networks can be estimated at significantly less than 20% of populated areas in the EU

Yet the industry has talked about 5G NSA vs. SA for at least 5 years, during which it became clear that moving to SA was going to be a huge challenge. While having enough spectrum for advanced services is indeed important, the real obstacle has been the creation of the cloud and software platforms needed for 5G SA networks, and the human skill and organisational changes required. (The UK set a target for 5G SA in its wireless strategy in early 2023).

Not only is has it proven very hard for MNOs to build the new “cloud-native” 5G cores, they have also needed to link them to other policy-related cloud initiatives such as edge computing and Open RAN. As a result, many telcos have needed to work more closely with cloud hyperscalers and platform providers such as Amazon AWS, Microsoft Azure, Google Cloud, Red Hat and VMware than in the past.

It is far from clear that consolidation of MNOs – especially during the same period this huge IT and network effort was ongoing – would have accelerated the deployment of 5G SA. It would likely have just added to the friction. It is conspicuous that greenfield private 5G networks have been adopting standalone at a much quicker rate, as they had less legacy infrastructure with which to integrate.

Yet rather than the white paper suggesting mechanisms such as better network/cloud training programmes or perhaps funding testing and interoperability labs, it instead suggests regulating those same cloud providers that are helping Europe’s MNOs reach their goal of fully-capable 5G networks that might actually drive demand and uptake.

Harmonising EU 5G/6G spectrum policy is unrealistic

The white paper also makes reference to the 5G “pioneer” spectrum bands intended for Europe, and notes that the process started in 2015 and has not yet been completed. It omits to mention that this included the Commission mandating release of the 26GHz band which has (predictably) proven to be almost irrelevant for 5G until now. This, together with the egregious original omission of including any spectrum for Private 5G networks, underscore the unsuitability of Brussels centralising coordination of spectrum policy. The recent Letta report which references the 6GHz band compounds this failure, as I noted in a recent post.

“In this context it should be considered whether, to ensure that new technology advancements are rolled out across the EU at the same time, an EU spectrum roadmap towards 6G should be enshrined in the law and enforced in a coordinated way by all Member States”

“This means reconsidering the role of the European Conference of Postal and Telecommunications Administrations (CEPT) in EU decision making, given the representation of non-EU Member States in this international body”

This is a spectacularly bad idea, and one which got a huge amount of pushback at the recent EU spectrum management conference in Brussels I attended. Even the EU’s own spectrum advisory group (RSPG) disagreed[5], once you got through the legalistic “due consideration” and “taken in to utmost account of” verbiage in its response.

That opinion clearly has a “but” and it’s the section after that which sticks the knife in with “the European spectrum harmonisation model, including the preparation of technical harmonisation of spectrum by CEPT, and the well-established cooperation with ETSI enabling standardisation, should be preserved”.

In short, spectrum is too important for nation states, they all have different national contexts, and the current EU + CEPT approach works well. It ain’t broke, so don’t fix it.

Conclusion and recommendations

The European Commission has consistently overestimated the potential benefits of 5G, and the timelines and practical limitations for achieving them.

Its white paper has many flaws, and its elements on 5G are no better than it’s ignorance of cloud/telecoms “convergence”, edge-cloud, interconnect or private networks. It gives two great examples of correlation being different to causation, when it attempts to assert that delayed spectrum auctions meant that “As a consequence, today Europe is lagging behind its international competitors on the uptake of 5G.” and then “However, while the single market, on average, delivered on price, it did not deliver on the mass deployment of advanced infrastructures and services like 5G standalone or the proliferation of advanced industrial and IoT services”.

Delays to 5G standalone and advanced industrial / IoT use-cases in Europe cannot be attributed to failures of the single market, or to spectrum non-harmonisation. They can instead be attributed to a five-year delay in understanding the importance of 5G private networks in enterprises, and continued efforts to push-back against efforts by some NRAs (such as in Germany, Finland and France) to offer spectrum to help accelerate their deployment.

The white paper should have shifted to a more expansive view of “Advanced Wireless”. It should cover not just public and private 5G, but also satellite-based networks for better rural coverage, Wi-Fi and dedicated in-building wireless systems, specialised IoT technologies and shared/hybrid networks and neutral host business models.

But rather than encourage the creation of a diverse wireless ecosystem capable of addressing all needs from city-centre basements to offshore windfarms, it has bought into GSMA, 3GPP and ETNO’s vision of a 5G monoculture of both technology and business model. It is allowing its own political desires to turn MNOs into “orchestrators” and to wrest control of spectrum policy from NRAs and CEPT to defy the reality of the marketplace and users’ needs.

To make a 5G-related analogy, this is a bit like the difference between AR & VR.

In virtual reality, you create a purely imaginary world, defining the characters – and even the physics – as you see fit. Conversely, in an augmented view, you add to or enhance the genuine external reality, as it exists in the real world.

If the Commission continues to fail to acknowledge its past errors and misjudgements on 5G, and tries to create its own reality without reference to real-world factors, it will just guarantee Europe’s “virtualisation” in the 6G era beyond 2030.

?

?

[1] https://5gobservatory.eu/observatory-overview/observatory-reports/

[2] https://digital-strategy.ec.europa.eu/en/library/2023-report-state-digital-decade

[3] www.dhirubhai.net/posts/deanbubley_5g-indoorwireless-wifi-activity-7185696231849963521-v_pA

[4] digital-strategy.ec.europa.eu/en/library/implementing-decision-setting-out-key-performance-indicators-measure-progress-towards-digital

[5] https://radio-spectrum-policy-group.ec.europa.eu/document/download/15789389-828c-4a55-a397-f2338fa2125b_en?filename=RSPG24-019final-RSPG_Opinion_on_how_to_master_Europes_digital_infrastructure_needs.pdf