Your browser is not supported. Please update it.

11 February 2019

As first generation 5G looks conservative, talk of 6G inevitably begins

Most leading mobile operators, when interviewed, make similar pronouncements about 5G:

  • it is more than a radio upgrade;
  • it needs to support brand new revenue streams, especially in the enterprise and IoT;
  • it is about convergence with fixed and non-3GPP wireless technologies, to create a ‘network of networks’;
  • it will only deliver all its commercial and socio-economic goals if it is deployed in parallel with a far broader overhaul of the network platform, from end to end, with the introduction of virtualization, software-defined networking (SDN), and networks that are disaggregated into a granular set of mix-and-match components.

It is dispiriting, if not altogether surprising, that the rush to stick a 5G flag in the ground means early movers are mainly deploying 5G, initially at least, in a very conventional way, and with very few clear links to any virtualization and digitalization roadmaps. To make matters worse, as if to compensate for the uninspiring nature of first-phase 5G, the industry is already starting to talk about ‘6G’.

Two communities – the vendors and academia, both of which have an interest in pushing ahead to new technologies – are starting to discuss what 6G might look like. The main tangible research is being done in two areas – very high frequency spectrum, including terahertz, and ultra-massive MIMO, with 1,000 or more simultaneous beams.

Bluesky research is always to be welcomed – it encourages new thinking and pushes the boundaries, and some of that trickles down even into current technology developments. It is when the 6G initiatives talk about their goals that things start to sound horribly familiar. The aim is to support very large numbers of IoT devices; to enable immersive experiences including the Tactile Internet; to create a ‘network of networks’ that might include better optical/wireless interfaces and the integration of satellite with terrestrial wireless. Surely these ambitions for a multi-technology network that could support new use cases enabled by VR, low latency, IoT and so on are exactly the same that were set out for 5G?

So on the surface, it seems that real world deployers of 5G have largely given up on the more challenging visions set out by governments and academics, and are focusing on familiar architectures and services – and as a result, most of the visionary activity has shifted to a ‘6G’ agenda. Of course it is a lot more complex. Whether or not 6G becomes a reality outside the universities and labs in future, 5G will have multiple phases. With echoes of 3G – which was deeply underwhelming in its initial phase but became very capable with HSPA – the first wave of 5G looks largely like 4G-plus. Same mobile broadband use cases; similar network architectures with very little virtualization or disaggregation – and therefore, the same challenges for the operators, in terms of having to invest to improve services for a mobile base which does not want to pay any more for the improvements.

Even one of the defining features of 5G, the use of high frequency spectrum, will not come into play in most markets until the 2020s, Verizon being a notable exception (see separate item). Most early adopters of millimeter wave spectrum – even AT&T or Korea Telecom – are not placing it at the heart of their deployment yet, because of the issues with using it for high mobility and wide coverage.

But all this will change while 5G is very much an evolving platform, and long before 6G is real. Release 16 will, it is assumed, fill the gaping gaps left in the radio specs so far. The decision to focus Release 15 mainly on MBB – which meant sticking to enhancements of existing OFDM waveforms rather than designing anything radically new – has led to many of the misconceptions about 5G, especially that it would deliver ultra-low latency and massive IoT support from day one. These applications, along with some of the immersive Internet use cases, will be no better enabled by 5G New Radio in Release 15 than by LTE-Advanced. But if Release 16 lives up to its promise, it will adopt additional waveforms geared to the IoT applications, driving towards true ultra-low latency for services such as advanced virtual reality.

No doubt, time pressures and political bickering among the 3GPP members will lead to some of the work anticipated for Release 16 being postponed until later. Already, many candidate work items have been removed from the Release 16 agenda – some will vanish without trace, others will resurface later. Integration with satellite was one example (and the satellite industry association has now joined 3GPP, presumably to improve the chances of getting it back on the table in 2020-2021). Others managed to get into Rel16 despite fears they would slip, 5G-Unlicensed being a notable one (and that was thanks to the sponsorship by 3GPP heavyweight Qualcomm).

When Releases 16 and 17 are taken into account, 5G should be able to support many of the visions and use cases identified for it – though operators and other deployers will still have to work out the revenue models for many of these, of course. At that point, the capabilities envisaged by early 6G research look like just an extension of 5G – and if radical new waveforms do find their way into the specs at the end of this year, in Rel16, there will be a strong argument that 5G should be the last ‘G’. The industry, finally equipped with a set of standards which addresses a huge range of network behaviors, applications and spectrum bands, could avoid the expense and disruption of another forklift upgrade. Instead, it could focus on leveraging technologies like software-defined radio and virtualized, multi-RAT networks, to create an ever-more flexible platform which could be enhanced in a modular, componentized manner – in the manner already seen in the cloud world – and so adapt easily as new use cases emerge.

That nirvana will only be real if operators plan and deploy the new architectures which will enable their 5G networks to support hundreds of new use cases and user experiences. A fully virtualized 5G core, eventually a virtualized RAN; transformation in the OSS and the transport layer; open networks and agile orchestration of all the virtual and physical elements – all these are critical to the business case, and much of the work on these is not being done by the 3GPP, or even ETSI, but by the new open, and open source, initiatives hosted by the Linux Foundation or Telecom Infra Project. These, or their successors, will be as important in the coming decade as the 3GPP – which should, for ‘6G’, focus on evolving the radio, not reinventing it, to move up into terahertz spectrum (if that is really required as early as 2030) with massive antenna arrays and so push towards terabit speeds.