Mobile World Congress, given its location, might traditionally be dominated by the European operators, but even European giants like Nokia were acknowledging that 5G will be driven, in its early years, by other regions, notably the USA, China and south east Asia.
In the wake of the Winter Olympics in South Korea, KT was able to claim the prize for the first commercial 5G service, even if it was limited in time and space to a single event and was pre-standard, and came with some very mixed reports about actual performance in Pyeongchang.
Nevertheless, KT claimed its system, which consisted of 22 5G links over 10 sites, delivered 3,800 terabytes of data during the two-week games, and reached peak speeds of 3.5Gbps on a Samsung 5G tablet.
“It’s one of the most broad-scale testbeds you can have,” said Aicha Evans from Intel, one of KT’s deployment partners. “We learned a lot.” KT built a pre-standard 5G network in 28 GHz spectrum, working with Samsung, Ericsson, Intel and others.
with vendors including Samsung and Ericsson for the Winter Olympics. Intel provided the FlexRAN and Mobile Edge platforms.
The four main use cases in the showcase were:
- a system that used hundreds of cameras to capture ice skating events and broadcast them in real time to virtual reality viewers
- connected Toyota automobiles
- transmission of high resolution pictures of cross-country skiers
- a fleet of 300 drones flying in formation during the opening and closing ceremonies.
Intel also talked about the 5G services it would help to deliver for NTT Docomo at the 2020 Summer Olympics in Tokyo, Japan. These will include 360-degree, 8K video streams, smart city services including pervasive facial recognition, and AI-enabled eHealth applications for athletes for training. Docomo says it will have launched a full commercial 5G network in Tokyo in time for the games and claims to be working with 600 different partners on the network and services.
Despite the news from east Asia, the race for first commercial 5G deployments is most intense in the USA, where even Sprint – which until recently was staying aloof, claiming its 2.5 GHz LTE roll-out would satisfy its users’ requirements for some years – is now plotting its path to the new networks.
Nicola Palmer, Verizon’s chief network engineer and head of wireless networks, got in first in the latest round of tussles with AT&T for the 5G crown, telling a FierceWireless event at MWC: “We’re going to be first to market”. At the same event, Gordon Mansfield, AT&T’s VP of RAN and device design, proclaimed: “We have every intention of being first.”
As AT&T has said before, it will get to “full” mobile 5G first, even if its rival is first to switch on fixed wireless services. AT&T has another justification for its claim – it will use full 3GPP 5G New Radio (NR) standards, while Verizon will initially use its pre-standard implementation and upgrade this later to be fully aligned with 3GPP. The existence of two phases of 5G NR standards is largely down to AT&T’s desire to be able to launch in the same timeframe as Verizon (rather than a year behind it, as in 4G), but to avoid the risk and expense of an inhouse pre-standard technology. It led the group of operators which pressurized 3GPP into fast-tracking a subset of the specifications, 5G NR Non-Standalone (NSA), which still requires a 4G core.
Some regard NSA as almost a US-specific technology, created for the convenience of AT&T, and many operators say they would prefer to be in less haste, and wait to deploy the 5G RAN and core together – Orange, for example, has been vocal about the trade-offs of investing in an NSA network, whose performance and flexibility will be limited by the 4G core; and then having to upgrade later to Standalone (those standards should be finalized in mid-2018).
China Mobile expects to be the first operator to deploy a Standalone network at large scale, in 2020, while some other early adopters, such as SK Telecom and NTT Docomo, will go with NSA in the first phase, in order to hit ambitious deadlines. The Chinese giant will run field trials in Shanghai, Hangzhou and three other cities from the second quarter of this year, with at least 100 base stations deployed in each. It will also run smaller scale trials in a dozen other cities, including Beijing and Shenzhen. Most of the trials will run in 3.5 GHz, with some in 4.9 GHz. Initially the equipment will be pre-standard and will be software-upgraded once the Standalone specifications are complete and equipment available.
A GSMA Intelligence report last year noted that the economies of a standalone deployment would be “significant” for a national roll-out, but these would be achieved only after the MNO has made a major upfront investment, since Standalone “is likely to prove more expensive at least in the early stages”. While the headlines often focus on the early adopters, the majority of MNOs (almost 80%, according to a recent Rethink survey) are prepared to wait until Standalone is fully available, the economics have become clearer, the spectrum allocated, and there are plenty of devices – and in the meantime, make do with enhancing LTE.
But there will be no such pragmatism and caution in the USA. AT&T says it will launch standards-based 5G in about a dozen cities before the end of this year. These will include parts of Atlanta, Dallas and Waco, Texas.
Verizon will launch a fixed 5G service in three to five cities this year, and Palmer said it will move on to a mobile offering in 2019.
Verizon has already tested its pre-standard V5GTF technology (which is very close to the NSA standard) in 11 markets across 200 nodes, and says it has been able to achieve gigabit speeds over fixed connections of up to 2,000 feet, despite the limited range of the millimeter wave spectrum it is using. In the USA – but again, not in most markets – 5G and mmWave, mainly 28 GHz and 39 GHz, have become almost synonymous. In other countries, like China, there is far more emphasis on midband spectrum like the 3.5 GHz C-Band (which in the US is currently being used for LTE).
Meanwhile, Sprint CTO John Saw said the operator would add 5G-capable equipment to “tens of thousands” of its towers during the next year, but would wait until smartphones are ready to launch wide-scale commercial services. “We’re aiming for an iconic-type phone,” he said. In the meantime, Sprint will bulk up 4G with densification and Massive MIMO, taking advantage of an increased capex budget of between $5bn and $6bn in 2018. Having spent only $2bn last year, it is now back in the same range as T-Mobile USA, which projects capex spend of $5.3bn. AT&T expects to spend up to $25bn this year in capex, led by fiber and 5G.
Saw showed off Sprint’s 64×64 Massive MIMO radio/antenna unit, supplied by Samsung, which will initially support 4G and then later 5G, which will be activated, in the 5G-ready towers, via a software upgrade. Sprint is working on technology to allocate 2.5 GHz spectrum dynamically between 4G and 5G depending on changing usage patterns. The first cities to get the Massive MIMO LTE update, later this year, will be Chicago, Dallas, Los Angeles, Atlanta, Houston and Washington DC, and these are also likely to be among Sprint’s first 5G cities from 2019.
As for T-Mobile USA, it promised to start building networks supporting 5G in both high and low bands in 30 cities this year, but would switch on commercial services only as devices became available, probably starting in 2019. CTO Neville Ray said the MNO will deploy 5G in 600 MHz, 28 GHz and 39 GHz spectrum in cities including New York, Los Angeles, Las Vegas and Dallas. The build-out will be split about evenly between Nokia and Ericsson.
The 600 MHz deployment has already started with 5G-ready LTE upgrades in the field but Ray was more cautious about mmWave, saying there was still “a lot to learn” before its precise potential was clear. He sideswiped at AT&T, saying: “Why are we in New York and not Waco? Because New York matters.” TMO has previously been scathing about the FWA-first 5G model adopted by the big two carriers.
In Europe, TMO’s parent Deutsche Telekom will start commercial 5G trials this year in “all its big integrated countries” and go live in 2020 in at least some of them, according to Claudia Nemat, board member for technology and innovation. The first trials will be in Germany, followed by Hungary and Slovakia.
Deutsche Telekom said on its recent earnings call that it would spend €5.7bn ($7bn) in capex in Germany this year, up from €5.4bn ($6.6bn) in 2017, led by fiber and densification as well as the start of 5G. Across all its markets, it plans to spend €12.5bn ($15.4bn) this year, compared with €12.1bn ($14.9bn) in 2017.
However, while vendors may be rubbing their hands in glee amid all these big capex figures, they should not get too eager. For one thing, operators – especially in Europe – are increasingly insistent that they will use the shift to 5G and virtualized platforms to shake up their supply chains and rely on a wider range of vendors, many of them with low cost solutions (see separate item). For another, many of the big capex investments are on fiber or IT. Operators have also been insistent that 5G must be justified initially on cost reduction, and while it may involve a couple of years of raised spending, it will need to be significantly cheaper than 4G was, both for deployment and total cost of ownership (TCO).
NTT Docomo of Japan, one of the 5G frontrunners, is not forecasting any capex uptick from 5G. Seizo Onoe, the operator’s chief technology architect, told an NGMN event at MWC that one of 5G’s “killer services” could be “data capacity enhancement with no increasing trend in capex”. The company’s capex has fallen from a peak of nearly one trillion yen ($9.3bn, at today’s exchange) in 2001, when 3G services were launched, to less than JPY600bn ($5.6bn) last year.
“In the past capex has been on a downward trend,” said Onoe. “We can continue this trend” by using open source and white box networks.
“Revenue is not growing and has been very flat or rising just 1%, and so there is no room for increasing costs,” said Johan Wibergh, CTO of Vodafone Group, at the same sessions. He has been one of the most vocal about the need to justify 5G first on cost reduction. “If you look at spectrum auctions, that is an extra cost. We don’t have more money to spend but less. That is the reality,” he added.
Deutsche Telekom has also indicated that overall capex is likely to fall next year, from a peak of €12.5bn ($15.3bn) this year, helped by initiatives like the newly formed ORAN (see separate item), which is targeting a white box, virtualized RAN.