UK regulator Ofcom has released its proposals for the next tranche of spectrum auctions. It has given the industry until March to respond to a call for consultation.
First, the basics – what is up for “grabs” in terms of spectrum. Ofcom will award 60 MHz of paired 700 MHz spectrum, 20 MHz of downlink-only 700 MHz spectrum, and 120 MHz of 3.6-3.8 GHz spectrum.
Maths geniuses will have quickly noted that comes to a total 200MHz. This additional spectrum represents an increase of around 22% of current overall spectrum holdings. The 700 MHz spectrum comes to 47% of all currently held low frequency spectrum, including 1400 MHz.
Who is likely to bid for what?
Obviously this isn’t up to Ofcom, but it does state the following.
Indeed it states: “We consider that EE/BT might well have a high intrinsic value for both 700 MHz (as it has little low frequency spectrum at present) and 3.6-3.8 GHz spectrum, which it could use to deploy a 5G service.
“We consider it highly likely that O2, Vodafone and BT/EE would all have a high intrinsic value for some 3.6-3.8 GHz spectrum, and that all four MNOs may well value some 700 MHz spectrum – either because they have little low frequency spectrum at present (BT/EE and H3G) or because they would use it to provide a 5G coverage layer in the shorter term.”
Competition and spectrum caps
Ofcom has decided that 37% is the maximum amount of total available licensed spectrum that any one MNO should hold. That means that EE/BT, the company with the largest amount of spectrum, will only be allowed to gain a maximum of 120MHz of the 200MHz total up for grabs. Vodafone could acquire 190 MHz and H3G 185 MHz. O2 would not be restricted by the proposed cap.
Ofcom says that it is not worried that any operator currently faces an existential threat due to an overall lack of spectrum. “We consider that O2 would have sufficient spectrum to remain a credible operator if it won no spectrum in this auction.”
Nor does it think that a situation where one operator has a lot of a certain band but not much or none of another band is necessarily a threat to market competitiveness.
The regulator also assesses that all the operators have a long term path to roll out 5G using current spectrum, even if they won none of the spectrum on the block in this auction. That’s partly because it expects, in the long term, all spectrum will come to be available for 5G. “All MNOs currently have some low frequency spectrum as well as at least 80 MHz of mid frequency spectrum. They therefore all have a long-term path that will allow them to offer a wide range of 5G services, even if they fail to acquire any spectrum in this auction,”
So it is sticking with the overall 37% spectrum cap it decided upon at the 2018 auction of 2.3 and 3.4 GHz auction, and is leaving it at that.
Contiguous spectrum in the mid band
Ofcom realises that spectrum that is contiguous, or at the least proximate, is more useful to operators than spectrum that is split apart within a band. That’s because it makes things like antenna configuration and RF planning easier. For example, you cannot currently aggregate two separate carriers in the 3.4-3.8GHz band, and such aggregation will not be supported in early 5G devices. And it is not possible to deploy a single active antenna system for separate carriers if the radio equipment at the base station does not cover the necessary frequency range.
“Active antenna systems are relatively large and heavy (compared to passive antennas). It may therefore be costlier or not possible for MNOs to deploy separate antenna systems on each mast for multiple spectrum holdings in the 3.4-3.8 GHz band due to issues related to space, weight, wind-load and power. This problem may potentially be exacerbated by the presence of network sharing agreements”
“Operators want to deploy 5G on carriers of 80 to 100 MHz. So those with 40 or 50 MHz in the 3.4-3.6 GHz band (BT/EE, O2 and Vodafone) are therefore likely to want to acquire additional holdings of 30-50 MHz in the 3.6-3.8 GHz band in this award. That would mean they come to hold spectrum in two non-contiguous blocks within 3.4-3.8 GHz.”
However, Ofcom judges that technology will massage some of these problems away, for example by enabling greater carrier aggregation within devices, or by operators using small cells to boost capacity in areas where active antennas are deployed in a smaller carrier.
Ofcom also says that, actually, some MNOs may indeed obtain spectrum that is close enough to their 3.4 GHz holdings to fall within the range of a single active antenna unit.
The regulator also thinks that operators could consider trading their holdings in order to get contiguous or proximate blocks.
In fact, some operators were keen enough to get contiguous blocks in the whole 3.4-3.8 band that they suggested the 3.4-3.6 GHz blocks that operators already hold should be included in the assignment stage of the 3.6-3.8 Ghz auction. Essentially this would mean reassigning spectrum (some of it already held) within the entire 3.4-3.8 GHz band to create contiguous blocks.
Ofcom rejected this idea, saying it doesn’t have the power to make those who have already got 3.4-3.6GHz spectrum put those bands back into the mix.
It also considered saying to Three that if it wanted to enter the new auction it would have to give up part or all of its holdings as part of that process. It didn’t think Three would go for that (!) and added that Three would probably see more value in holding onto that spectrum to trade it later.
In fact, it thinks it likely that Three would trade, this is because there are likely to be benefits from wider defragmentation of the band, and “H3G will only be able to profit from this if it trades its current location in the 3.4 GHz band”. Also, it may trade some of its mid band spectrum for some sub 1GHz from another operator.
So how will Ofcom enable trading? Will it be post assignment or part of the assignment process. Consider this issue, one operator gets 20MHz between two bands held by another operator. It could hold the second operator “hostage” by preventing it from having a contiguous band. On the other hand, it too suffers if it has other holdings in the band.
Perhaps, then, Ofcom could restrict any smaller band awards to the top or bottom of a band, thereby stopping a small band sitting in the middle of otherwise contiguous spectrum? However, this could result in a bidder not getting the 40 GHz band it bid for at the top or bottom of a band, because that has been reserved for a smaller band. Another approach might be to allow contingent bidding, whereby an operator says, “We bid this for this band, but only if that operator gets the band next to us, so we can make a trade.”
In the end, it looks like Ofcom is going to go for a conventional assignment stage (ie not contingent bidding etc) and then look for some sort of brief period of trading to see if operators can work things out after.
“Our view is that allowing a negotiation window after assignment stage bidding may be the most straightforward option for facilitating post auction trades through the auction, without introducing potentially material implementation or bidding complexity.”
In essence, Ofcom recognises the problems of non-contiguous spectrum between 3.4-3.8 GHz – a problem that critics say it created by releasing the band in two stages – but is going to leave it up to the market to deal with those problems, via technical improvements and by post-award spectrum trading.
Coverage obligations and incentives – and a missing 2%
The regulator is proposing offering a discount on license fees for operators that make a commitment to boost rural and building coverage. It wants to get good outdoor coverage up to 90% of the UK landmass – which pretty much means only leaving largish chunks of Scotland (about 25%) and Wales (17%) uncovered.
It says that it estimates operators would already have about 82% coverage by the time of the award, and thinks there are significant benefits to getting to 90%. However, it says that its initial target of 92% begins to see costs outweigh the benefits. Therefore 90% is the number it now likes.
It estimates that a “typical” operator with around 82% coverage could increase its geographic coverage to 90% by deploying new sites for approximately £200-400m. However, revised analysis suggests that the incremental cost of increasing coverage from 90% to 92% (which is the level of coverage that Ofcom initially proposed) is £110m-£140m.
These coverage incentives for the new 90% target would require that, within four years of the award, the discount license holders will: “deliver good quality 4G mobile coverage outdoors to at least 90% of the UK landmass; deploy a minimum of 500 new wide area coverage sites; and provide new coverage in areas where at least 140,000 premises are located.”
Note that an operator may use any mobile spectrum to deliver this obligation, so that an operator wishing to bid for 3.6-3.8 GHz, but not 700 MHz, could still bid for a coverage obligation if it wished – and deliver it using alternative spectrum bands, e.g. 800 MHz.
Nor does Ofcom think it has all the answers to getting those last few percent rolled out. It says that license terms are only one lever that authorities can pull, with others including easier planning processes, new rules on repeaters, public funding and new rules on sharing and roaming.
Sharing spectrum for low power, local private networks
In a separate document, Ofcom released a consultation into new approaches to enabling spectrum sharing and re-use. At first this will see sharing of spectrum on a “static” basis, ie – not the dynamic spectrum sharing enabled by the databases that enable the likes of CBRS or TV Whitespace networks.
The idea here is to see users own local networks, and also in some cases to use bands that are licensed to mobile firms but not currently being used.
Third parties (ie not MNOs) that operate at low or medium power in shared bands could open private networks for industrial IoT applications, extend rural broadband connectivity using fixed wireless access (FWA), or urban and rural not-spots (indoors and outdoors). They would also need roaming agreements with national operators to allow “local” users to stay connected when and if they travel outside of the local area.
Ofcom is proposing to enable shared access in three bands. The 3.8-4.2 GH band is currently used by satellite Earth stations, point-to-point fixed links and wireless access applications (fixed) by UK Broadband. Deployments in the band are technically coordinated by Ofcom on a first come, first served basis. Ofcom says that the band could be used for private networks that automate processes in a range of industries and more broadly support IoT. It is adjacent to the 3.4-3.8 GHz band which has been identified as a primary 5G band in Europe. 5G technology standards cover this band and radio chipsets supporting this band are available for equipment vendors to develop equipment.
At 1800 MHz shared spectrum the 1781.7-1785 MHz paired with 1876.7-1880 MHz band is currently authorised to 12 Concurrent Spectrum Access (CSA) licensees on a shared basis through an award process. Ofcom is considering expanding access to the band to a wider set of users and also administrating that access under the same scheme it proposes for the two other new shared bands. This band could be used by third parties to extend mobile coverage in rural areas and in buildings, or for local private mobile networks.
The 2390-2400 MHz band is the spectrum between the recently awarded 2350-2390 MHz mobile spectrum and licence exempt uses above 2.4 GHz. It currently supports some MOD deployments. As above, this band is already supported by 4G mobile networks and handsets making it suitable for similar applications as in 1800 MHz shared band but with more capacity.
Licenses in these spectrum bands would see low power and medium power operations.
A low power licence for local connectivity would allow users to deploy as many base stations as they like within a 50 metre radius circle without further authorisation from Ofcom. Potential licensees could apply for multiple adjacent licence areas if the required coverage area is larger than the area defined by a single licence. Outdoor low power base stations will be limited to a maximum height of 10 metres above ground level. This is in order to reduce the risk of interference to other new users of the same channel. The maximum EIRP that Ofcom proposes for low power base stations is 24 dBm per carrier for carriers up to 20 MHz or 18 dBm / 5 MHz for larger bandwidth carriers. In addition, it proposes that the medium power base stations support a fixed service only. The maximum EIRP that we propose for these is 42 dBm per carrier up to a 20 MHz carrier or 36 dBm / 5 MHz for larger bandwidth carriers.
The cost to local, private operators would be £80 per 10 MHz (for 3.8-4.2 GHz and 2300 MHz shared spectrum, so 20 MHz = £160; 100 MHz = £800 and £80 for 2×3.3 MHz (for 1800 MHz shared spectrum).
Additionally, Ofcom said that it would like to see evolution towards Dynamic Spectrum Access (DSA), but it thinks its proposals are the quickest to get low power, local networks to market.
“We believe our proposals would provide the quickest route for new users to access the proposed three shared access bands. We note that a DSA approach could provide users more flexible access to spectrum as devices would automatically connect to a central database and be assigned spectrum based on availability at that time and location. From a spectrum management perspective, this would also ensure an efficient use of shared spectrum but a DSA solution would likely take longer to develop and test. We will, however, explore the potential for introducing DSA in the three shared access bands. We would like to work with industry to define the appropriate specification for both DSA equipment and database capability that would enable future transition to DSA.”
“To facilitate the evolution towards DSA, we consider there is benefit in creating an industry group to discuss both the technical requirement of the DSA equipment and database to inform a future regulatory framework for DSA.”