Is DSS a game-changer for the 5G business model?

Two main pushes on the network side. One surprising, the other less so.

It was a week to split the mobile network commentariat. In Barcelona, Qualcomm held its annual 5G Summit. In Switzerland, Huawei re-upped its Mobile Broadband Forum. In The Hague, speakers at NFV-SDN World were grappling with cloud native operations, automation and softwarisation of the network. And down the road in Amsterdam the rather more fixed-network focussed Broadband World Forum assembled to discuss whatever. 

Your correspondent, not yet being able to overcome space-time limitations and subdivide himself in 5G holographic form was only able to attend one of these. And so… to Barcelona.

This event is a big meeting opportunity for Qualcomm. It brings in operators and “ecosystem partners” – the device and equipment manufacturers, the cloud players.  

It’s instructive to consider what Qualcomm itself is pushing to its audience. The things where Qualcomm gets all rigid in its message discipline are usually either the things in which it thinks it has a lead on its competitors or where it stands to gain a bunch if operators/equiment/device vendors deploy. Not every bet it places will come off, but it doesn’t need them all to. In the past this might have meant noise about things like LAA and MulteFire, of multi component carrier aggregation (think #GigabitLTE anyone) and the like. If the ecosystem thinks these things are a good idea, economically viable and provide a potential competitive advantage – ultimately that’s more wins for Qualcomm.

So from a network technology point of view here’s what Qualcomm was primarily dealing this week; mmWave and Dynamic Spectrum Sharing.


The message here was that you really need mmWave to get the “full benefits” of 5G. Secondly,  mmWave also works indoors and works well. 

Qualcomm has long been a mmWave proponent, even two years ago conducting coverage simulations to prove that you can get decent outdoor coverage even from a 4G grid (as long as it’s fairly dense to start with). The point then was: you don’t necessarily need tens of thousands of new sites for the small cells that will make mmWave work.  

The point now is that by deploying mmWave you not only get the bandwidth and low latency for the most extreme applications, but you get an overall capacity uplift across all the spectrum layers of 5G. Cristiano Amon, Qualcomm’s President, said that “real” 5G needs all the bands, and is in fact essential for the advanced sorts of use cases such as industrial automation and the like. And he told journalists at a round table event that getting mmWave to work indoors is no problem.

Amon: “Inside it propagates very well; the only thing it doesn’t do well is outside to inside. Inside it is going to be deployed like WiFi enterprise or retail APs and will be intrinsic to realise connected enterprise and industrial use cases.”

Amon name-checked uses such as cloud based storage and processing for connected devices, on-demand computing, mixed reality devices and control of automated robots. “That’s TerraBytes of info at very low latency – you cannot realise that without mmWave.” 

Picture from Neil Shah

Of course, in an arena, factory, large room, anywhere with a notable absence of walls, the first part of this statement is certainly true. mmWave antennas are not going to find many obstructions. But start throwing in staircases, lift shafts, interior load-bearing walls and the like, and it’s not so easy. Not that Qualcomm’s engineers would necessarily agree. They think that a couple of things come to the operators’ aid. One is using the reflective paths and natural blocking of walls to extend and control propagation. Another is the use of advanced beam management to keep massive MIMO beams attached to devices. Meanwhile CoMP (Co-ordinated MultiPoint) can be deployed to send signal simultaneously from multiple access points, without incurring re-transmissions.

Qualcomm’s position that mmWave is all good for indoors actually puts it at odds even with others who are selling mmWave solutions. It’s a notable point of difference as it insists mmWave designs will be suited to enterprise, industrial and other indoor venues.

Staying with mmWave – Rakuten’s Tareq Aminsaid that he thinks Japan could have the largest mmWave deployments globally in 2020. That’s because Japan has a lot of lamp poles, and it’s easy to get on them. Already the inter-site distance in Tokyo is about 200m, and with more poles Amin thinks he could get it down to 50. Amin is pretty confident about the in-building enterprise opportunity for mmWave as well, and he also likes the FWA or in-home opportunity. He just needs to get to a window and the CPE can do the rest, making it a self-install solution. 

Rakuten is working with Airspan for its 28GHz antenna-radio units (build on a Qualcomm FSM100xx chip), connected to the vBBUs it runs from Altiostar. So if Amin is serious about this level of deployment then that could be very good news for Airspan. The chipset vendor also has a reference design for mmWave small cells.

Qualcomm 5G small cell reference design with FSM100xx chipset, mmWave RF Module (Picture via Ignacio Contreras )


The second network message was seemingly a bit more esoteric. Qualcomm went all in on Dynamic Spectrum Sharing (DSS). Why?

DSS enables 5G NR and LTE signals to exist in the same spectrum band – the FDD lower bands that LTE operators in – effectively scheduling LTE transmissions in the gaps that exist in 5G NR transmissions. 

DSS is a R15 technology, but it was a late-breaking one. It is also an Ericsson-developed technology that was pushed forward through 3GPP by Ericsson, giving Ericsson a natural lead in the technology (and an IPR advantage??). In fact, the technology nearly came a cropper quite late in the day as a bug was identified that would have impacted performance. This blog outlines how “with the help of device chipset vendors” (wonder who that could be referring to) a technical workaround was quickly achieved. 

Ericsson’s head of 5G commercialisation Thomas Noren said that DSS entails being able to compute on a millisecond basis on the radio baseband – essentially shuffling the signals in amongst each other quick enough to reduce any performance impact.

Noren told TMN that some PRBs (Physical Resource Blocks – the frames that make up a subcarrier) are “lost” in DSS mode – but that downside is more than compensated for by the upside of increased 5G coverage. The spin is that by using DSS along with Carrier Aggregation you improve the overall uplink capacity, and that in turn increases the cell size of the higher band site, because coverage is mostly limited by uplink capacity.  

Noren: “This is very complicated scheduling to do. If you go back 18 months one of the other vendors said they had it and then they realised they didn’t, so then then they said, this is impossible. And now they say they will also have it again. So I think many thought that this was a very challenging thing to do but we have proven that it is possible.

“The others should speak for themselves but it seems like we have a very powerful baseband. This enables us to handle this kind of scheduling and avoids those overlaps in PRBs. I think this is an advantage for us. I’m sure others will do this but they have to answer, with which chipset have you tested this, can you do it in the frequency domain and can you do it per millisecond. If you cannot answer positively to all three, then I think you have an issue.” 

DSS, then, is a way to keep the business case for 5G NR afloat, by allowing operators to re-use existing spectrum for 5G. By doing that, its advocates say, you effectively extend the coverage of the higher band as well. There is still, however, the whiff of some smoke about the actual trade-offs in terms of performance, and what headspace is required for control signals within the LTE and 5G NR channels.

But let us come back to Qualcomm. Why does it care about DSS? Well, here’s the thing. Cristiano Amon, President of Qualcomm, said at the event that the devices are ready for 5G. There’s a lot of them. So there can be “no excuses” from operators to delay 5G rollout because they are waiting for devices. 

However, one possible valid hold-up is a lack of spectrum that would easily support widescale coverage. Most of the new awards are in midband spectrum, which is not great for wide area rural and indoor penetration, and/or in mmWave, which is useless at that.

And many of the new sub 1GHz bands – like 700MHz – are not yet on tap. So that means that a natural thing to do might be to refarm or otherwise reuse your existing bands. The problem is, these bands are full of 4G customers who are providing good business, and you aren’t going to switch them all up to 5G any time soon, no matter how much you might like to. But what if you could deploy 5G NR in that 4G spectrum and also leave your 4G network alone, how would that look? 

Ericsson and Qualcomm – and already Swisscom and Verizon – think that would look pretty good. Qualcomm’s X55 modem supports the feature and Amon said to be on the lookout for more devices that support DSS at MWC Barcelona 2020. So it may have been a surprising point of focus, but it is one that Qualcomm is positioning as a “game changer” for the 5G rollout model.

Disclosure: Qualcomm paid for TMN to travel to this event.