The most recent meeting of 3GPP’s RAN working group in Dubrovnik, Croatia did not just rubber-stamp the AT&T-driven proposal to have a two-stage release of the initial 5G New Radio (NR) standards.
The decision to accelerate progress by releasing a non-standalone specification (which still needs an LTE anchor network) ahead of the full standard was the most attention-grabbing development. But while controversy rages over that approach – with critics claiming 3GPP has sacrificed quality under pressure for a quick result- there was also quieter progress on other aspects of the specs which will make up Release 15 and even its successor, Release 16.
Much of the excitement around all phases of the 5G standards revolves around new spectrum, especially above 20 GHz. But while trials and technologies are mounting up, especially in the 28 GHz band, there is still deep uncertainty in some key areas, including:
What will be the profitable business cases, and will they be fixed or mobile?
Where will WiFi, which has its own millimeter wave technology in 60 GHz, fit into the 5G mix?
And do most operators really need the high frequency spectrum at all, when many have large amounts of underused airwaves already, and should be concentrating on new ways to monetize those?
mmWave 5G technology gathers momentum, but what is the business case?
One of the earliest schisms in the 5G world is over the most immediate business case for the new networks, especially when they are deployed in high frequency spectrum. Verizon and AT&T are among the leaders in trialling bands such as 28 GHz, and are mainly focusing on fixed wireless.
This has the advantage that there is no need to wait for a broad base of attractive handsets to build up before services are adopted, and for these two telcos specifically, there is the opportunity to fill gaps in the fixed broadband network and extend out of their wireline territories. Verizon has said that the fixed wireless case can justify 5G build-out on its own, but rival T-Mobile USA is less sure (and, of course, doesn’t have a fixed broadband business).
Its CTO Neville Ray said he was “not a huge believer” in the fixed wireless model, telling a recent investor conference in London: “We’ve done a huge volume of 5G testing, we have 5G radio, 28 GHz, all those pieces.” But he echoed the sentiments of other operators such as MTS of Russia and Telstra of Australia in questioning how immediate is the business case. “Every business case piece of work I’ve seen put in front of me doesn’t math at this point in time,” he said, citing the cost of high density deployment and the propagation challenges in 28 GHz.
He believes the attractive use cases for mmWave spectrum will be related to virtual reality, artificial intelligence and other emerging technologies, and harnessed to mobility, but this convergence will take years to achieve. “I think there’s a plethora of really exciting 5G stuff that will come, I just don’t think [fixed] is a great place to start,” he added.
He also doubts that mmWave bands are the place to start, supporting the view of many that, just like 4G, 5G will be deployed in low frequency macrocells initially, and then capacity will be added via higher bands and ever-smaller cells. He managed to criticize yet another rival in making this point, telling the Citi event: “I think Sprint has this phrase that 2.5 GHz is the new low band of 5G. Complete BS. 5G is going to move across all spectrum bands over time. You cannot cover the US with 2.5 GHz spectrum.”
On the other side of the world, in Japan and Korea, there is also high interest in mmWave spectrum, but a greater focus on the even greater challenges of mobility in these short range frequencies. Softbank of Japan, working with Ericsson, has announced one of the most comprehensive trials to date in 28 GHz, encompassing indoor and outdoor connectivity, both fixed and mobile. Along with 4.5 GHz, 28 GHz is expected to be a leading 5G band in Japan.
The Tokyo trial will use Ericsson’s mmWave 28 GHz 5G Test Bed solution, which includes prototype base stations and devices and will support Massive MIMO, Massive Beamforming, Distributed MIMO, Multiuser MIMO and beam tracking.
Last November, Japanese mobile market leader NTT Docomo completed a 5G trial with Samsung in 28 GHz. This achieved a data speed of more than 2.5Gbps in a vehicle traveling at 150 kilometers per hour.
Whether fixed or mobile, Ray believes – and is borne out by many studies, including Rethink’s own operator surveys – that outside the US and a few hothouse Asian markets like Japan, the need for huge additional doses of spectrum capacity is not yet urgent, certainly not enough to justify the difficulties of deploying a network in very high bands.
A new study – by Finnish consultancy Rewheel and Digital Fuel Monitor – suggests that operators have large amounts of underused spectrum, before they need to return to the auction fray or move to 5G and mmWave. This should be plenty to support the return to unlimited data, or to expand into the quad play, for years to come.
The report claims that operators can move into fixed wireless broadband with existing spectrum, especially in the unpaired 2.5 GHz/2.6 GHz band.
The analysts explained: “We took operators’ spectrum holdings and the number of macro sites, applied generic assumptions for traffic geo- and busy hour distribution and sector spectral efficiency and ran our model with operators’ 2016 data volumes. We repeated the model for a scenario where the operators’ existing TDD spectrum in the 2.6 GHz band was also put into use. Then we ran the model again for a Massive MIMO scenario. Finally we ran it for a scenario where new spectrum is added in the 2.3 GHz and 3.4-3-8 GHz TDD bands (with Massive MIMO).”
They found that most European mobile operators used only a tiny percentage of their available capacity; and that they could support 100GB per person per month right now, while if they added TDD and Massive MIMO, they would be able to support 200GB (or 500GB per household) – more than enough for wireline substitution, for those carriers which do not have fiber resources, but are looking to compete in the quad play arena.
Substitution is already happening in some markets, says Rewheel, including its native Finland, where a rising percentage of people are moving away from home WiFi and doing everything, including fixed access, on cellular, as low cost unlimited data plans once more become the norm.
Some start-ups expanding the fixed wireless ecosystem:
At the IEEE event in San Francisco, several start-ups where showing off their innovations for millimeter wave spectrum.
Cambridge Communication Systems (CCS) showed a compact fixed wireless system which is already being used by Softbank of Japan and sold by Ericsson and ZTE. It delivers 1.2Gbps over distances of around 500 meters, in the spectrum from 24 GHz to 29 GHz. The company is seeking a new round of financing to fund development of a 10Gbps version and build support for 28 GHz and 60 GHz bands.
Tarana Wireless also showed a fixed wireless system, this one for bands from 3.6 GHz to 5.8 GHz. It uses 16 antennas on either end to handle non-line-of-sight and delivers 200Mbps over one kilometer, running proprietary protocols on an FPGA.
Blue Danube has its antenna technology in trials with several operators including the AT&T fixed wireless project. In LTE, it supports 96 antennas per base station with beamforming but no requirement for a remote radio head, in order to boost throughput and signal quality at the same price as conventional base station antennas.
Phazr has developed a low cost fixed wireless platform based on a WiFi ASIC and incorporating 384 antennas in the base station and 64 in the CPE. It promises data rates of up to 30Gbps over one kilometer with line of sight, or 200-800 meters without, in bands from 24 GHz to 40 GHz. The company says a UK service provider has already tested the technology and CableLabs plans a field trial later this month. Phazr is backed by FiberTower, now being acquired by AT&T.
Artemis demonstrated the new compact version of its two-year-old pCell technology, which claims to create a personal cell of coverage around a user on-demand, allowing for highly flexible and efficient use of spectrum and network resources. The new miniature cell, which was announced last year as a far smaller implementation of pCell, is initially targeted at home use. It is based on the same 10-mm2 FPGA as Google’s Webpass and incorporates nine antennas to aggregate 100Mbps throughput from a service running in the 600 MHz band.
National Instruments said it conducted the world’s first public demonstration at 28 GHz of a real time, over-the-air prototype that aligns with Verizon’s 5G specification. The OFDM system used eight component carriers in a 2×2 downlink MU-MIMO configuration with hybrid beamforming and a self-contained subframe, achieving 5Gbps peak throughput. NI says it is scalable to over 20Gbps with eight MIMO streams.