Verizon says it is on track to migrate from its pre-standard ‘5G’ radio specifications to full 3GPP compliance, but has no plans to attempt 5G mobility when it makes its first deployments in 2018.
The US operator has been trialling 5G fixed wireless access (FWA) and plans to offer commercial services in some areas from next year, as does AT&T. The two rivals took contrasting approaches to getting ahead of the curve with 5G.
Verizon worked with a group of vendors (its5G Technology Forum, which includes Cisco, Ericsson, Intel, Qualcomm and Samsung) to define specs to enable it to deploy before standards were finalized, while AT&T used its influence and partners to push 3GPP into fast-tracking a subset of its Release 15 specs so some initial standards would be ready at the start of 2018 (the 5G New Radio Non-Standalone, which requires an LTE host network as the 5G Core will not be ready).
Verizon introduced its inhouse specification, 5GTF, in July 2016 and has now run tests in 11 cities mainly in the 28 GHz band, working with Qualcomm, Samsung, Ericsson and others. These began, in the first quarter, in business and residential districts of cities in New Jersey, Massachusetts, Texas, Washington DC and Michigan. Then they have been expanded to new metro areas including Ann Arbor, Atlanta, Bernardsville (NJ), Brockton (MA), Dallas, Denver, Houston, Miami, Sacramento, Seattle.
There has been some concern that Verizon might create fragmentation in 5G standards by developing its own specs, but it is already planning an over-the-air trial using the full 3GPP New Radio in 2018. Those tests, like the pre-standard trials which the carrier has already started conducting, will take place in 28 GHz and 39 GHz bands.
In a statement, Verizon and Qualcomm said they “plan on delivering a common 5G NR mmWave technology platform for mobile and home broadband wireless access, supporting a 5G NR migration path for Verizon’s early 5G fixed wireless access deployments and trials based on the 5G Technology Forum specifications.”
There are very few details as yet about the actual migration process and how much effort or expense that will involve, but the risk is considerably less in a fixed wireless environment, since there is less need for interoperability and smartphone support is not a factor. The same process was seen in 4G, when some fixed wireless operators which had deployed WiMAX eventually moved to TD-LTE, but were able to make that change when it made economic sense rather than in a desperate haste to align with standards.
Once operators want to support mobile services, that desperation sets in, since they need to be fully in line with the norm to enable roaming and to access the best devices. NTT Docomo learned that lesson in the 3G era, when it raced ahead of the industry to deploy its own pre-standard FOMA technology in 2001, in order to steal a march on its rivals. In those days, it had sufficient control over its inner circle of handset makers, including several locally based ones like Sony and Panasonic, that it was able to secure attractive and affordable devices. But as Japanese consumers started to want iPhones and the open mobile Internet rather than a carefully curated Docomo experience, the disadvantage of slightly off-beam technology and spectrum became more glaring, and was also clear in roaming deals.
By 2009, FOMA was in use by 91% of Docomo’s base with over 50m subscribers, but it had left the operator over-dependent on Japanese handset makers, with the high costs of slightly off-standard devices; and with networks that were starting to age and even become obsolete.
That led Docomo to push into LTE early, but in order to re-enter the mainstream and start phasing out creaking FOMA equipment, rather than to set its own de facto standards. Then-CEO Ryuji Yamada said in 2009: “We went first and when we looked behind us there was nobody there…[This time] we don’t want to go first, we want to go in the leading group.”
Dan Warren, director of technology at the GSM Association at the time, also highlighted the risks of going pre-standard, saying in an interview: “NTT has equipment in its network that is very obsolete … every upgrade to 3G is more painful than the last, while Softbank has hit the point where its equipment is good enough that it can continue upgrading 3G relatively cheaply.”
NTT Docomo invested heavily in R&D for 4G and 5G but did not decide to take a pre-standard route again after the lessons of the 2000s. AT&T will have been aware of this cautionary tale when it pushed to accelerate 3GPP efforts rather than emulate Verizon and develop its own technology. This is partly because AT&T has a less compelling motivation to major on FWA. Its fixed-line footprint is far larger than Verizon’s, and while it sees FWA as useful to replace ageing DSL or penetrate some of its rival’s north-eastern territory, the business case looks far more marginal than for Verizon, which could potentially harness 5G to deliver a wireless-only quad play in regions where it cannot offer fiber.
So for AT&T, FWA is a convenient way to test 5G in anger at the earliest stage, particularly to assess the real world performance of mmWave spectrum. It is easier to launch fixed services in the early days because there is no need to cajole the device makers into launching high profile smartphones, still the main driver of consumer uptake of mobile services. But mobility is central to the first-wave case for AT&T.
Verizon clarified last week that, despite its reference to mobility in its statement about 3GPP migration, that would not be supported using its 5GTF equipment – in other words, it will wait at least until 2019, and for fully standardized networks. CFO Matt Ellis said on the carrier’s third quarter earnings call that mobile 5G was “certainly not a 2018 activity”.
Qualcomm has said it expects early 5G smartphones to appear from some point in 2019, but these are unlikely to support mmWave bands, but will likely prioritize on lower and mid-band 5G spectrum such as 3.5 GHz (especially in Asia).
AT&T and KT have both suggested they want to introduce some mobile services in 2018, but they would presumably have to commission their own devices, which would be costly and would need to be justified by strong confidence in the impact of first mover advantage. In reality, no operator will have broad coverage with 5G until 2020 at the earliest, and so mobile users will still be on 4G most of the time, shifting to 5G in certain downtown areas, venues or other hotzones. T-Mobile USA has said it aims to begin mobile 5G roll-outs in 2019 and has been largely dismissive of the FWA case.
But Verizon remains committed to it, and Ellis offered updates on the progress of its FWA trials on the earnings call. He said there would be a full progress report issued later in this quarter, but there had been interesting results from the tests, including “
the “fact that we can deliver service without needing line of sight”; and the ability to deliver services to apartment buildings more than 20 floors in height, which was previously considered infeasible.
According to Ellis, Verizon will continue – like most 5G frontrunners – to expand the LTE network with new spectrum and with densification, as this will remain the “workhorse” for the initial 5G era. Just over half of Verizon’s mid- and low-band spectrum is currently in use for LTE and additional airwaves in 850 MHz, PCS and AWS bands will be activated during 2018. Ellis also said Verizon would continue to deploy more fiber as the backbone for multiple services including 4G densification, 5G, Internet of Things and residential broadband.