AT&T has revealed that Dynamic Spectrum Sharing (DSS) is now live in portions of its network to facilitate migration from 4G to 5G network coverage.
The emphasis is perhaps on the word ‘portion’ at this stage as AT&T, like its US counterparts, is proceeding towards 5G in a piecemeal fashion, introducing it in areas of denser population where an enhanced experience is likely to gain most appreciation.
AT&T has been deploying 5G using millimeter wave spectrum for high capacity in a few selected areas, rather disingenuously calling that 5G+. Like Verizon, AT&T has limited midband 5G spectrum, at least until the auctions in the CBRS and C-Bands, though it has started to extend coverage using its 600 MHz assets – but admits these give a scarcely visible performance improvement over 4G because of the limited capacity in the band.
The need for midband spectrum with a balance of capacity and coverage has driven high interest in DSS with the two largest US operators. DSS provides a temporary stopgap, and a stepping stone from 4G to 5G that avoids the need for full refarming of spectrum. Static re-allocation risks either proceeding too slowly, with delays to full support for 5G services; or too quickly, in which case the rump of 4G users are starved of capacity.
AT&T was coy about which vendor it has been working with, but it must be Ericsson because Huawei was the only other supplier with a fully commercially available DSS product at the time AT&T started to deploy.
Ericsson has been leading the DSS charge and been promoting it almost as its own by calling it ‘Ericsson Spectrum Sharing’. Huawei has closed in – though not in the US market of course – and also claims to have successfully developed and tested 1ms spectrum switching, considered essential for effective DSS operation (see below).
AT&T’s other network suppliers, Nokia and Samsung, promise to deliver the capability soon in the USA, and Nokia is already supporting DSS in some networks, including working with Vodafone Idea in India. Nokia has worked hard to close the gap on Ericsson, having claimed deliveries of its DSS software began in April 2020 with volume shipments slated for July.
China’s ZTE has lagged behind Huawei but with a similar approach, promoting it as an extension of its Magic Radio Solution, while Samsung has been more muted.
On the operator side, the USA has been in the vanguard, with Verizon making the early running with a lot of talk during 2019 and then demonstrations with Qualcomm and Ericsson in November that year. Since then, however, Verizon has been rather vague over precise roll-out times, without clarifying exactly what issues will determine full commercial availability. Like other operators it has been using Covid-19 as an excuse for delays and has only really confirmed that DSS remains part of its plans to establish nationwide 5G coverage by the end of 2020.
But as we have observed before, use of the word ‘nationwide’ to define cellular coverage is usually misleading, especially in the USA where large rural areas still do not even have 4G coverage.
When it comes to DSS, operators’ attitudes tend to be colored by their perceived need for the technology, often associated with shortage of spectrum in those low bands around and just below 1 GHz. Those like T-Mobile USA, that are relatively rich in coverage spectrum, and so are under less pressure to rely on sharing with 4G, tend to play down the benefits of DSS and accentuate the drawbacks. TMO’s president of technology, Neville Ray, has criticized DSS for being late, while arguing that compromises to get it working have chipped away at the capacity to the extent that the gains are less compelling.
There may be some merit in Ray’s argument that it is worth waiting for evidence that the software coming on-stream from Ericsson, Huawei and Nokia is working well and is stable, but the gains in flexibility and scalability from dynamic sharing will outweigh slight losses in aggregate bandwidth. Apart from scalability and flexibility, DSS will ease many MNOs through their migration by providing that layer that maximizes combined 4G and 5G coverage while the network is transitioned from Non-Standalone to Standalone mode. When that transition is complete, 4G can theoretically be switched off, though only when almost all subscribers have 5G-capable devices.
It is true there is some trade-off, since the monitoring and management involved add some extra traffic and possibly even latency. There is also a question over how much performance 5G will add over 4G in a low band, which depends on factors such as width of channels and cell density. But millimeter wave spectrum faces even greater challenges to deliver the promise of massively increased capacity. They include poor indoor penetration and need for far greater density to achieve sufficient spherical coverage for consistent performance at the higher frequencies, and also need for extensive testing of the devices with their higher levels of integration.
It will be some time before mmWave, or FR2, is widely deployed and meanwhile DSS will play a major role ushering in 5G not just in the USA, but around the world. Some European operators, such as VodafoneZiggo in The Netherlands along with both Sunrise and Swisscom in Switzerland, will actually roll out DSS ahead of the US telcos.
Indian operators are even more spectrum-starved than their US counterparts and do not have any 5G-specific airwaves yet, since auctions in 3.5 GHz have been postponed. Again, DSS may be a way to start deploying 5G before they have new spectrum.
The key to DSS lies in the degree of dynamism supported, for to deliver on its promise of effective sharing of spectrum in real time it must be able to reallocate it on the fly, almost instantaneously. For that there is agreement that 1ms scheduling must be enabled, so that spectrum can be allocated to 5G as required quickly enough to satisfy the more demanding use cases, including low latency, as well as network slicing generally. There is also the prospect of DSS living on well beyond 4G to 5G migration by facilitating coexistence with private or local services sharing the same spectrum (see separate item).
So DSS, either as a temporary migration step or for longer lasting coexistence, must underpin the more demanding aspects of 5G, subject to the constraints of the low spectral bands in which it operates. Ericsson has attempted to seize this territory as its own, but Huawei has closed that gap to the extent it now lists 1ms DSS as one of its 10 key enablers for 5G.