Nokia's 4.9G races Ericsson's almost-5G, yet the finishing line is a mirage

4.5G 'Pro', 4.9G and 3.9G! Oh come on...

End of the line (train line). Photo by Shutterstock

At Mobile World Congress this year, Nokia emerged as the most aggressively "5G-ready" of the big vendors, while Huawei was still stressing "4.5G", a term it coined, but which has been widely adopted. This week saw Nokia going one better with "4.5G Pro" and even "4.9G", recalling the build-up to LTE, when some vendors and operators were still majoring on 3G extensions, and insisting on labelling HSPA+ "3.9G".

Mobile vendors score no points for originality in the buzzwords they apply to their new generations of network gear, but the endless reworking of the "Gs" as virtually meaningless marketing labels does, sadly, obscure real progress in real technology – and a genuine dilemma for OEMs.

Who really wants early-stage 5G?

As a new generation of standards approaches, vendors have a difficult balance to strike, between driving new sales of the current generation of kit, while trying to make customers invest in the next one at an early stage. In fact, in an ideal OEM world, there would be far less rush towards 5G. They would be able to keep selling updates to LTE for a great deal longer, with their costs falling all the time as it became a mass market platform. That would deliver a better return on the investment in 4G development and marketing, especially as each generation of equipment becomes cheaper in capital expenditure terms (a point NTT Docomo and others are making very pointedly in 5G conferences).

Likewise, Rethink’s biannual survey of almost 100 mobile network operators (MNOs) has consistently revealed that - beyond a few cutting edge operators which always need to push the envelope in network performance – most would prefer to squeeze as much life out of LTE as possible before embarking on upgrades with all the accompanying disruption and risk, and issues such as device availability to grapple with. And both the traditional vendors and most operators would like to have more time to get to grips with what 5G might be, especially since it is likely to be heavily virtualized, which introduces a whole new set of migration and skills issues.

But several factors mean the vendors and operators will not get their way – 5G will become a commercial reality, in one form or another, by the end of the decade and there will be huge pressure to start adopting it, in parallel with 4G expansion in most cases.

Fierce competition among the network OEMs has hardly dissipated following the consolidation of Nokia and Alcatel-Lucent, because there are new players like Cisco and HPE riding the new virtualized architectures to take a bite out of the mobile network pie, and even the established suppliers face slowdown in key international markets. Operators and vendors alike are under pressure from shareholders to be ahead in 5G, and find it easier to get support for a new product if it can be labelled 5G, or at least seem like a firm step towards that.

So, reluctantly or not, the whole community is being dragged towards a 5G which is not defined or, in many cases, really wanted yet. For most operators, the new iterations of LTE combined with moves towards virtualization and SDN (software-defined networking) would be more than enough transformation for the next decade, but now they need to implement those changes within a framework which can be migrated to 5G.

This gives vendors the chance to keep selling more LTE equipment while portraying this as a future-proof route to 5G. Ericsson played that hand strongly in June when it announced its "plug-in" software modules, which promised to support key aspects of the modern network, such as advanced MIMO and virtualization, via modules which could be implemented in 4G systems but would be "5G-ready".

Pre-standard claims wear thin

This entails the tried and tested "xxx-ready" approach – if you buy equipment now, it will improve your networks and your economics immediately, and will then be smoothly upgradeable to the next generation when that is available. In the past, this has usually turned out to be a false promise. One of the WiMAX industry’s fatal mistakes was to pretend that all kinds of proprietary broadband wireless systems were "WiMAX-ready" when they were not, and then to introduce a second WiMAX standard, 802.16e, and pretend that it was an easy matter to upgrade, in software, from the earlier, incompatible one. Similarly, promises of simple transitions from 3G to LTE were illusory.

The "pre-standard" argument only really works when that standard is so close that only minor adjustments are likely to be made – Wi-Fi equipment is often released ahead of the final ratification, with few, if any, modifications needed to be fully compliant.

The 5G situation is somewhat different. This time, vendors have to deliver, not just promise, a smooth transition from 4G to 5G, because operators will not tolerate another big bang upgrade. Instead, they need to introduce new capabilities one by one, according to their business requirement, so that a current HetNet gradually morphs into one with 5G characteristics.

Nokia’s stepping stones to 5G

This was the real story behind Huawei’s original announcements of 4.5G, and gave a marketing label to something most operators did want – extensions to 4G capabilities, but with a pre-5G label that would satisfy investors and some more savvy customers. Now Nokia has appropriated that term too, and has announced "4.5G Pro" while looking ahead to "4.9G" – presumably 3GPP Release 14, the last one before 5G (Release 15) kicks in.

Nokia will bring 4.5G Pro (a snappier name for LTE-Advanced Pro, in turn the commercial label for Releases 13 and 14) to market in its AirScale radio platform at the end of this year. It is promising “10 times the speeds of initial 4G networks”, and a seamless upgrade for current customers of Nokia 4.5G (LTE-Advanced). These number about 90 worldwide, says the company. Among the features which will power 4.5G Pro’s speed boost are five-band carrier aggregation.

Meanwhile, "4.9G" is being billed as a parallel technology to early 5G, allowing operators to continue to enhance their LTE networks even as they start to deploy the next generation standards (or at least, this is how vendors hope it will go. Some operators want to continue to squeeze the most out of 4G for many years before they move towards 5G).

According to Nokia, 4.9G will bring “significant capacity and data rate enhancements and network latency reductions to let users maintain a continuous 5G service experience complementing 5G radio coverage”.

It will boost speeds to “several gigabits per second”, by aggregating even larger numbers of carriers and using highly directional antennas which will “allow signals sent via multiple transmit / receive paths to be added together”. It also promises to enable cloud-based RANS, with significant intelligence at the edge to reduce latency to below 10ms. This is an area where Nokia has been a pioneer among the network OEMs, particularly via its work on Mobile Edge Computing with Intel.

But while Nokia’s cloud and MEC activities may be far more relevant to most operators today than actual 5G (and really do provide a stepping stone to the future network, whatever its air interface) the mobile vendors remain fascinated by the radio standards themselves. This is the area where they undoubtedly rule the roost, so while Nokia was talking about 4.9G, Ericsson was unashamedly promising fully fledged 5G.

Ericsson promises 5G in 2017

It has made the bold promise to deliver the first 5G networks in 2017, even though the 3GPP standards will not be finalized until 2018. Arun Bansal, head of the Network Products business unit, said the 5G hardware plus the plug-ins meant “the first operators can start to deploy 5G infrastructure”, while for those sticking with 4G for now, the company is also “launching innovations that improve both the performance and efficiency of today’s networks using concepts that will evolve into 5G”.

Ericsson certainly has as good an insight as any company into what 5G might be, arguably leading the field when it comes to labs and trials. But there were still many question marks surrounding its announcement of its “first 5G New Radio (NR) product”, the Air 6468. In fact, this can will be usable now for 4G deployments (in TDD only – an FDD version will follow) as soon as it is commercially available in 2017. It is designed to be combined with the existing Ericsson baseband and plug-ins, either for 4G or pre-5G.

The 3GPP has dubbed its 5G radio standards work New Radio, but will not start on the work of specifying that until March 2017 – currently it is going through the process of receiving submissions from around the industry, and there are still many arguments about whether a brand new air interface will be needed for NR, or whether 5G will require different air interfaces for different types of spectrum (OFDM in sub-6 GHz and something new in millimetre wave, for instance).

So it is somewhat premature for Ericsson to be saying that its new radio, together with plug-ins including Massive MIMO, amounts to a 5G platform when combined with the vendor’s existing radio system baseband, the 5216, which is also used in its Radio Test Bed. The new product has 64 x 64 MIMO antennas to deliver 5-6 times more capacity than current RANs. Initially, it will be available in unpaired 2.6 GHz spectrum but "over time it will support all frequencies below 6GHz”.

Of course, just like the marketing labels, jumping the standards gun is predictable. Some vendors and operators are always so keen to drive, rather than follow, the market that they will launch prestandard platforms despite the risk of dead ends for themselves and fragmentation for the broader market.

Verizon drives its own 5G specs

The leading US operators are already engaged in a spat about this, with AT&T claiming that Verizon’s trials of "5G" in high frequency spectrum are proprietary and will therefore fragment the market. And the promised 5G networks in South Korea and Japan, for the 2018 and 2020 Olympics, are almost certain still to be running on pre-standard gear – though Japan’s NTT Docomo will have learned some lessons from being the first mover in 3G, when it rolled out the semi-proprietary "Freedom of Mobile Multimedia Access" services and found itself in a cul-de-sac which took years to exit.

Despite those experiences in the past, some MNOs remain undeterred. Apparently ignoring AT&T’s taunts, Verizon says it believes the 3GPP specifications are “75 per cent to 80 per cent there … at least for a fixed wireless use case”, according to strategy director Sanyogita Shamsunder. The operator has a 5G Technical Forum in which it is devising its own platforms with many industry partners, and it told LightReading last week: “We expect that the vast majority of this ends up in the 3GPP specification, with some software changes down the road.”

Because of such operator interest – and gun-jumping – Joakim Sorelius, product manager at Ericsson’s Business Unit Radio, believes there is a real need to accelerate matters this time around. He told Mobile Europe: "This is more of a personal reflection but it's fair to say we have seen movement in 3GPP to accelerate everything from being ready for the end of 2019 to it being moved forward."

He thinks initial standards will be published in the early part of 2018, and insisted that his firm is determined to follow standards as they evolve and not create a technology fork.

He said: "We are heavily involved with 3GPP and have had field trials and research in place since 2014. We are quite confident we are very familiar with the key technology concepts and these radios can support them with a software package. We are fully committed to the 3GPP version of standards. The radios have been designed with that in mind."

MIMO is particularly prominent in Ericsson’s current definition of 5G. The company says the Air 6468 will be the first to support Massive MIMO and Multiuser MIMO and is designed to integrate advanced antennas with a large number of steerable ports to enable many MIMO approaches and advanced beamforming.

Ericsson expands 4G portfolio too

Ericsson’s 5G announcement was part of a bigger launch of radio equipment. Indeed, most of the launch was focused on enhancements to the flagship RAN platform, the Ericsson Radio System, to “target the peak productivity years for LTE technology”. These enhancements included three gigabit LTE radios equipped with 4x4 MIMO. The Radio 4407 and 4412 enable 4x4 MIMO in a single radio unit for FDD and TDD respectively, while the Radio 8808 supports advanced TDD beamforming applications. The firm also focused on some other key aspects of the HetNet - LTE small cells for unlicensed spectrum, and baseband units supporting flexible site builds, to enable dense rollouts.

The Radio 2205 supports unlicensed spectrum in small cells, while a pair of new baseband units - Baseband 6502 and macro Baseband 6303 - are designed to meet the growing need for densification. Ericsson also unveiled the Baseband P614, which it said addresses interference in dense network builds by enabling new bands to be activated; and the Baseband C608 for Cloud-RAN.

As well as actual products, it launched a network deployment delivery platform, which provides a single-visit site verification process spanning installation, configuration, integration, shakedown and handover of the site, to speed up rollout in a dense environment. And it added a Spectrum Analyzer, a software product designed to identify external interference without the need for a site visit. This shows Ericsson incorporating the key elements of modern HetNets and, presumably, of 5G into its portfolio at a steady rate. It has been a little behind the curve on virtualized RAN, compared to Nokia or Huawei, but has been adding those capabilities this year. Its latest launches reflect its belated, but now intense, interest in small cells and densification, which dates to the Radio Dot début. The combination of licensed and unlicensed spectrum is also important to LTE and 5G, though the NR product – like the 3GPP’s initial deliberations – is focused on traditional mobile bands, not the millimetre wave frequencies in which the US operators are putting so much faith for their early stage rollouts.

Copyright © 2016, Wireless Watch

Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here.




Biting the hand that feeds IT © 1998–2018