5G is 'ready' once you redefine 'ready'... and then redefine 'reality'

It'll be big one day. But that day is not tomorrow

Analysis If the industry had one job at Mobile World Congress last week, it was to tell the world that 5G – the biggest thing since "electricity or the automobile", according to Qualcomm's CEO* – was almost upon us.

But the reality on the ground in Barcelona only emphasised how far away it remains, practically, for most of us. From network deployments to handset readiness, the word that doesn't spring to mind is "imminent". Unless your hobby is testing networking kit, it's not going to be worthy of your attention until the second half of 2020.

This isn't to say 5G isn't going to be very interesting. That's because 5G is really an umbrella term that covers so many new technologies and opens up so many new business use cases. Under the umbrella is almost every theoretical use case, business model and bleeding edge labs innovation any one has dreamt up in the past decade.

But for the next 18 months, it's largely going to be a debugging exercise. Two coruscating expert summaries underlined this.

One report (PDF), from D2D analyst Jay Goldberg (formerly of Lazard, Deutsche Bank and Qualcomm), noted how "operators are very nervous about rolling out 5G". So much is new or not yet working.

With consumer benefits not exactly imminent, the industry has naturally focussed on industrial and enterprise use cases. Goldberg found "a dizzying array of science fiction future applications powered by the wireless network. We heard that one operator presented 600 use cases to their equipment providers last year."

But few actually require 5G.

He added: "During the show we tried to count use cases. Unfortunately, many of the use cases failed the very basic question of 'Couldn't we do that with 4G?' In all we found a few dozen use cases that could credibly claim to be both dependent on 5G and actually useful."

Source: D2D

For Europeans, the picture is probably worse than Goldberg paints, because of the piecemeal spectrum picture, and national priorities. There's no pan-European network and no benefit to getting the bragging rights to being the first European 5G deployment (nobody cares).

Fewer than half of European countries are yet to hold their 5G spectrum auction, though the UK's Ofcom got the first stage out of the way last April and there's more to come.

Stretching the meaning of deployment

Last week we described how, for once, the handsets weren't as late as the network. Compared to 3G and 4G, more handsets will be on tap at "launch". But these things are relative. Technology remains immature at both ends: the RAN (radio access network) and the "terminal".

Edison Research's Richard Windsor filed some devastating posts from Barcelona that exposed the bluster. Windsor called it "a disappointing situation which also did not hold up well to close inspection or questioning". Only one of seven handset vendors was demonstrating a true mobile 5G handset. The rest were mockups, hidden, had radio disabled, or were sending data via another transport, such as Wi-Fi or a cable.

Windsor did note, however, that the breakthrough millimetre wave technology had some surprises.

Of the one that worked, he said: "The mmWave antenna at 28Ghz is 12" from the terminal but at least one knows that it works as placing a hand in between the two stops the video."

So that's all right, then.

Live 5G calls were arranged for the media's benefit, but similarly failed the reality test. The "audio quality [was] from the 1950s and latency as if the other guy [were] on the Moon," tweeted one blogger.

You can see the immense potential for 5G. For many it will eventually supersede DSL and cable. We got a glimpse of that with HTC's 5G Hub. We'll eventually see cheap low-power local "campus" 5G networks. And network slicing is a huge leap forward for networks. You can almost forgive the Qualcomm CEO's hyperbole. But the latter two are really years away.

What the industry has done is declare a milestone for the consumer-facing bit of the spec, eMBB, which allows deployments to proceed. But it's misleading, as we explained here. Much of the rest has yet to be finalised.

To make us believe that it's "ready", the mobile industry has subtly redefined what we understand by the words "deployment" and "launch". When Hutchison threw the UK's first 3G network (Three UK) open in spring 2003**, it was at least national. And when EE launched 4G, it ramped up pretty rapidly.

The newly branded MNO first rolled out LTE in 10 cities, but added six more within a month, and hit 35 locations within around four months. Within a year it boasted 98 per cent population coverage. But then for 4G, the UK was a relative laggard – we weren't debugging the kit. And the speed, for mobile warriors keen to be free of Wi-Fi, was a blisteringly fast improvement over 3G (at least at first). Dongle owners were delighted.

It has been quite a reality check. Before MWC, some pundits forecast trouble for Apple because it probably wouldn't have a 5G iPhone this year (Apple will use Intel modems because of its dispute with Qualcomm). But it can probably safely skip 2020 too, at this rate.

Stand by for the longest public debugging exercise in history. ®

Bootnotes

* "5G will have an impact similar to the introduction of electricity or the automobile, affecting entire economies and benefiting entire societies," he said at CES 2017. Yes.

** Readers with long memories may recall Three launched on 3 March i.e. 03-03-03. Geddit?