Qualcomm took well over 2,000 staff, partners, customers, operators, journalists and other types to Hong Kong this week for its 4G5G Summit. You can read our report of its key 5G announcements here – but take from them the principal point that the event was all about making 5G a commercial reality by 2019.
So, with that in mind, here are 9 things we learnt at the event this week that will help us understand more about 5G.
1. Qualcomm is really really keen on some recent simulations it has done that show that operators will be able to get pretty decent coverage from mmWave access points without having to find, equip and deploy a whole load of new sites. It says that RF propagation simulations have shown operators could get 75% coverage just by deploying mmWave access points at sites that already host LTE base stations. The assumptions were of a reasonably dense grid in the first place, in urban environments.
The coverage statement was repeated many times at the event (TMN first heard it last month at RANWorld) – so there’s no doubt it was way up Qualcomm’s messaging grid for the event. It’s clear that Qualcomm is trying to head off concerns that 5G mmWave will necessarily require very small cell sizes, and therefore vastly more cell sites. Even aside from Qualcomm’s presence in developing the network side of things, if Qualcomm thinks it has a big lead in enabling mmWave at the device level, it makes sense that it doesn’t want operators to be scared off rollouts because of logistical concerns.
Qualcomm wants to see operators embrace mobile mmWave, giving it the opportunity to leverage its advantage at the device level. It’s also going to have a lot of IP in the infrastructure layer, thanks to its R&D in mmWave RF behaviour.
All about the average throughput increase
2. There’s reasonable clarity on the 5G-ness of 5G. In other words, what will be the things that actually makes the 5G air interface 5G? For Durga Malladi, SVP Engineering, this included:
– an optimised OFDM waveform in almost all cases save some specialised apps – for example RSMA for IoT use cases,
– ME-LDPC (Multi Edge Low Density Parity-Check) coding,
– a flexible sub frame design that enables multiple services on the same frequency,
– a scaleable numerology relevant to the subcarrier spacing of the frequency being supported.
– integrated backhaul and access
– Advanced Massive MIMO capability: that’s because it is M-MIMO that enables an overall uplift in average throughputs, and clearly increased cell edge performance. At mmWave bands this will require very fast beam switching within the same Access Point (as the user moves around) and between different Access Points.
The 5GNR smartphone reference design
3. Why does this matter? Well, we’re still in a pre-standards world, and Qualcomm is touting a 5GNR capable modem and supporting RF modules to support a smartphone-sized reference design. We’re not in FPGA any more, this is ASIC based. If there are changes coming to the air interface, then a hardware redesign would be… problematical. But Cristiano Amon, President of the chipset business at Qualcomm, answering this question, said, “Many of the aspects of hardware are consistent now and allow us to go forward. We will continue to see changes as the standard goes forward but at this point we are confident the X50 design will be R15-compliant to support Sub 6GHz and mmWave. We don’t see further steps as required features but to the software, and supporting increasing requests for combinations of frequency bands.”
4. Here’s another technical one: mobile mmWave has a very different power efficiency curve compared to mmWave WiFi. So Qualcomm is suggesting the higher the throughput – utilising a far greater bandwidth – the more efficient power usage actually gets. It’s all about overcoming objections to mmWave. Serge Willenegger added that new Power Amp and filter technology, and the adoption of antenna modules within the device design, should mean that operators get more aggressive and less conservative in their business models around mmWave.
Unlicensed spectrum and 5G
5. Qualcomm’s Dean Brenner, SVP Spectrum Strategy & Technology Policy spoke on spectrum. The gist was that all spectrum is available to 5G. All of it. Licensed, unlicensed, shared. This will enable private networks as well as licensed-operator networks. 5G in unlicensed standalone mode (the 5G equivalent of MulteFire) will be part of the specs, instead of being a 3GPP outsider. And other spectrum sharing techniques are also possible, beyond using timing mechanisms in Listen Before Talk. For example, spatial sharing could be enabled. Sharing will also be baked in, instead of being applied retrospectively as per LAA. Big implications for IoT and private, enterprise networks. Unlicensed specs will be worked through from R16, rather than in the first releases, however.
6. Don’t forget China. China is already entering Phase III trials of 5G technology. These government-led trials are now moving from proving out individual technology use cases to looking at more overall systems design and interaction from core to access layers, as well as the issue of transport. Phase III trials start from 2018.
7. Looking at massive MIMO in the 3.6MHz band might be a good way to go in terms of using rooftop sites, says Telstra’s Mike Wright. May be more efficient than getting down in the clutter at street level, with very small cells. Wright added that the industry needs to get better at explaining to each other why 5G is different. Where 5G can go that 4G doesn’t. Most of all we need customer use cases, Wright added.
With that in mind the operator will have a 5G slicing platform in place as the Gold Coast Commonwealth Games become an opportunity to try out some 5G use cases.
8. Not that many of the operators are actually fully on board with the 2019 schedule, despite the hype at the event. China Telecom sees deployment in 2020. But KT is one that will go early with some 2018 launches – to meet the deadline of its winter Olmypics. The company will have 5G connectivity through certain tunnels, is working on 360 degree video streamed live at certain sites using an edge cloud architecture, and is looking at a 5G bus which will have AR displays actually overlaid on the windows of the bus.
Customers must not break a leg falling between 5G and LTE.
9. What Qualcomm terms Gigabit LTE but you could just call LTE-A PRO, will be a key underpinning for 5G. You cannot break a leg falling from 5G to LTE, as Telstra’s Wright put it. So you need a really good LTE network as much as possible.
