Thank you for the opportunity to speak to you today.
Accommodating changing demand
Accommodating changing demand for spectrum is our core business. As Australia’s ‘town planner’ of spectrum, it is our job to identify changes in how spectrum will be used and ensure that spectrum is made available to meet the demand for new technologies and new services.
I don’t need to tell this audience that there are considerable and often competing demands for access to spectrum, and if anything, the pressure for changes seems to be speeding up.
Causes include the ongoing evolution and growth of the Internet, manifesting as continuing soaring demand for wireless broadband connectivity, and the more recent rollout of Internet of Things (IoT) networks and applications.
You are probably familiar with how the resultant ever increasing demand for wireless data has driven pressure for clearance of additional bands for terrestrial wireless broadband, fixed as well as mobile. While this shows no sign of abating, we are also seeing very rapid growth in innovation in the satellite industry, resulting from opportunities such as earth stations in motion, ‘smallsats’, looming ‘mega’ non-geostationary satellite constellations and ever-increasing throughput in traditional geostationary satellite networks.
The result is growing and changing demand for spectrum from a range of industries and users, including Australia’s innovative satellite and space industry, wireless broadband providers looking to implement 5G services. We’re also experiencing pressure from other existing spectrum users seeking to adapt and respond to all these changes, and from emerging new spectrum users, including utilities, the automotive industry, over-the-top service providers such as Facebook, and start-ups.
In recent years, the ACMA has released a regular five-year spectrum outlook or ‘FYSO’, which identifies both the emerging spectrum management challenges, and a detailed workplan for the coming year. While feedback on each year’s FYSO is a critical input to our spectrum management agenda, this year, for the first time, we have released a draft FYSO for consultation.
Reflecting on past feedback, we structured this year’s draft 2018–22 FYSO to include:
- a summary work program so that key work streams, activities and proposed timelines for engagement can be more easily referenced
- a holistic treatment of all major band planning and re-planning activities to support the establishment of new spectrum uses—this replaces the previous focus on replanning activities driven by the ACMA’s mobile broadband strategy
- more comprehensive attention to the wide set of existing spectrum uses, where annual planning activity is directed to optimising use within existing regulatory frameworks.
Thank you to everyone here who made a submission to the draft FYSO. We received over 30 submissions, and we are currently considering the comments that you provided on our work program priorities. You can expect to see the final 2018–22 FYSO published in this first quarter of the financial year.
So, against that general background, I’ve decided to focus today on two particular aspects that illustrate aspects of the dynamic nature of the sector and the differing impacts on demand for spectrum. They are satellite spectrum requirements, and wireless broadband developments.
Satellite spectrum
With growth and innovation in the provision of satellite-delivered telecommunication services and satellite usage in space science services, there is increasing pressure both internationally and domestically to ensure that regulatory arrangements continue to evolve to support this change.
For many years, spectrum planning efforts to support satellite services in Australia have focused on refining existing arrangements in bands currently used. A good example of this is the updating of legacy licensing arrangements to support earth stations in motion in the various parts of the Ka band (27.5–30 GHz) in recent years.
Innovations such as the deployment of low-cost, miniaturised space hardware (often referred to as nanosats, CubeSats or smallsats), the growth in the use of earth stations in motion in the fixed satellite service bands and the development of large constellation satellite systems are challenging the suitability of existing licensing and planning arrangements.
Demand from the satellite industry is growing. The satellite industry has indicated it may require greater access to the:
- L-band (around 1.5 GHz) for mobile satellite service (MSS) use
- S-band around 2 GHz to support mobility applications
- Ka-band (around 28 GHz) to support ubiquitous user earth stations beyond the approximately 1 GHz spectrum already available
- Q/V-band (around 40 GHz for downlinks and 50 GHz for uplinks) for feeder links for future high throughput satellite systems.
The short message is that we recognise and will respond to these demand pressures as our other priorities permit—I’ll just make a couple of quick observations before I invite you to wait for the final version of the FYSO.
In relation to the L-Band, we have already identified the need to investigate how additional spectrum can be used to complement existing ‘L-band’ mobile satellite service (MSS) allocations used by numerous satellite operators. As coexistence with potential wireless broadband use below 1518 MHz is likely to be an issue, a simultaneous review of the extended MSS L-band and the 1.5 GHz bands looks appropriate.
The S-band—2 GHz—was the subject of a number of industry submissions to the draft FYSO, not only from the satellite industry but from those interested in the band for television outside broadcast, terrestrial wireless broadband and direct air to ground communication to aircraft. What we should do next, and whether we can commence a public discussion this financial year, will be a key issue on my mind as we finalise the annual work program for 2018–19.
In addition to the investigation of bands for new satellite uses, but no less important, we have an ongoing program of work to optimise the planning arrangements for spectrum already used by satellite operators, including measures to better accommodate use by smallsats, earth stations in motion in the Ku-band, and updating our licensing procedures to support new uses.
Wireless broadband
Wireless broadband, including mobile broadband, has been a major driver for many spectrum management changes over the past two decades. ‘5G,’ the latest incarnation of wireless broadband, will see terrestrial wireless broadband pushing, for the first time, above 6 GHz into the so-called ‘millimetre waves’.
In response to the ever-widening spread of frequencies that are in demand, the draft 2018–22 FYSO has, for the first time, outlined different planning strategies for wireless broadband services in ‘low’, ‘mid’ and ‘high’ band spectrum. As they raise different planning and allocation challenges, I’d like to spend a little time on each of them.
Below 1 GHz
Sub-1 GHz spectrum has long been a bedrock of mobile communications, and while the epithet of ‘harbourside spectrum’ now gets applied to higher bands as well, particularly as 5G draws near, sub-1 GHz remains critical for its unique propagation characteristics. These include wide-area coverage and greater in-building penetration, and demand remains high relative to supply. Indeed, 5G itself gives a hint as to why we need to continually monitor and adjust these arrangements, because as 4G becomes, dare I say it, ‘old’ technology, we anticipate that attention will come back to re-farming lower bands for ‘coverage and/or control layers’ in the 5G architecture.
This means that sub-1 GHz planning arrangements are not ‘set and forget,’ and we need to be considering putting in place the building blocks for optimisation of the band for 5G. For example, anything we can do to enable aggregation of mobile spectrum into fewer, broader-band holdings where possible will minimise the amount of stitching that might need to be unpicked if (or when?) sub-1 GHz 5G gains momentum.
To respond to these pressures, we have been working with industry and government to optimise configuration of the 900 MHz ‘GSM’ band. As part of this optimisation plan, the ACMA also proposes the reallocation of the so-called 850 MHz ‘expansion band’ for wireless broadband.
Are there ways, in the long-term, we could make even better use of our precious sub-1 GHz wireless broadband allocations? Is there some way we could consolidate the bands themselves—potentially offering a ‘defragmentation dividend’? Some thinkers are imagining a world where the distinct sub-1 GHz bands such as 700, 850 and 900 MHz are restructured into essentially a single band, with the corresponding possibility of efficiency gains by removing guard bands and adopting a TDD structure as opposed to the traditional FDD approach now used. It’s a big idea, which raises many hard, practical questions, not least of which is how would it be done?
But that said, the idea has already been socialised in the Europe with the release of the paper entitled, The defragmentation dividend, A more efficient use of the UHF band. If you are interested in hearing more about this ‘big’ idea, Marc Eschenburg from Aetha consulting, which drafted the paper, will be speaking at the ACMA’s RadComms conference in late October.
As the holders of property-like rights, the initiation of broader trading to aggregate holdings such as this must be driven by licensees, but a sector hungry for new spectrum will be more effective in demonstrating the case for difficult changes in other bands when they can show that they are making the best use of their current holdings. As the regulator, we can help, or at least, get out of the way—by ensuring all sub-1 GHz wireless broadband allocations are in private hands in the form of fully tradable, long-duration licences; by aligning the tenure arrangements of these licences; and by enabling licence transfers.
If the key challenge for sub-1 GHz spectrum is to optimise existing allocations, this is not to rule out the possibility of additional wireless broadband spectrum becoming available over time. In addition to the allocation of the 850 MHz expansion band that I’ve just mentioned, the ACMA continues to monitor the 600 MHz band as a potential re-farming candidate.
The pressures on current arrangements in sub-1 GHz bands extend beyond wireless broadband demand. We’re also seeing the Internet of Things placing new pressures on class licensed spectrum at 900 MHz. In response, we are making additional class-licensed spectrum available to support low duty cycle, low power, devices optimised for wide-area Internet of Things networks. This new band, adjacent to the existing 900 MHz class-licensed band, will facilitate greater access for bespoke Internet of Things applications such as LORAWAN and Sigfox. This is an important reminder that the Internet of Things goes well beyond just 4G and 5G services.
Mid bands
Turning to ‘mid band’ for wireless broadband—1–6 GHz—an interesting reflection on the pace of change, is that not long ago, we would have called this ‘high band’ spectrum …
In the 1–6 GHz bands, the intensity of interest from wireless broadband has massively increased, in part because of 5G but also simply to meet the capacity needs of today’s market.
Globally, we are seeing other jurisdictions considering making a lot more spectrum in these bands available for wireless broadband uses, including in the broad 3.3–4.2 GHz ranges. Locally, our impending auction of 3.6 GHz spectrum will add to the substantial amount of spectrum already available in the nearby 3.4 GHz band, ensuring Australia can be a world leader in 5G deployments. We are also currently considering the priority of 1.5 GHz band replanning in light of submissions to the draft FYSO and will continue to monitor international developments in nearby spectrum in the 3.8 GHz band.
As is true with the sub-1GHz bands, the inefficiency of current wireless broadband allocations challenges us to optimise the configuration of the spectrum industry already holds. Recently, we’ve seen significant defragmentation of existing mid-range allocations at 1.8 GHz. Now we need to make better use of existing holdings in 3.4 and 3.5 GHz. As with 900 and 850 MHz, optimising these allocations will require action by both existing spectrum licensees, ACMA and government in the months and years ahead. The impending auction, and the minister’s recent direction to the ACMA concerning synchronisation, are vital first steps in this direction. Watch this space.
Also similar to sub-1 GHz, the 1–6 GHz bands have other incumbent users with substantial and often growing business needs. There is going to be a need for careful decisions over time to meet competing demands, including by exploring dynamic spectrum access or other sharing arrangements, where these can work.
Where changes are required, management of any disruption caused by such planning changes calls for continuing innovation and compromises on all sides. In the 3.6 GHz band, for example, we have sought to mitigate the disruption of re-farming through reallocation periods of unprecedented length—up to seven years. We have also paved the way for geographical sharing of the contested band, with proposed new satellite protection zones well away from the most heavily populated areas.
High bands
Now turning to the high bands, millimetre wave spectrum is the new frontier. The issues spectrum regulators face here are materially different, both in terms of incumbency (although it is not free of incumbents), but also in terms of the spectrum propagation and other use characteristics.
This presents an opportunity for the ACMA to think differently about licensing and ensure opportunities that are not just traditional fixed or mobile network operator use cases.
An immediate area of focus for the ACMA is the 26 GHz band (which extends from 24.25 to 27.5 GHz).
This band has been identified for 5G uses, having been included in relevant equipment standards and being the primary focus of 5G spectrum studies within the International Telecommunication Union’s current World Radiocommunications Conference study cycle. Those studies centre around compatibility with other systems that use the same (or nearby) spectrum and are important for identifying which frequencies will ultimately be harmonised internationally for 5G. Resolution of spectrum harmonisation studies is important to protect existing neighbouring uses but also to achieve economies of scale for network kit.
This international work, in particular the progress made with standards, has led to pressure in Australia and other countries to release spectrum in bands such as 26 GHz as soon as practicable.
Through the ACMA’s leadership, Australia has been heavily involved in these international studies, and I am pleased to say that progress has recently been made in European spectrum discussions, which will hopefully pave the way for agreement on harmonised frequency arrangements being considered in the International Telecommunication Union.
As some of you will already be aware, the uncertainty on some aspects of the international work on coexistence of 26 GHz wireless broadband with other services has unavoidably slowed down our domestic planning process a bit.
While we had hoped to be in a position to put forward a set of options for the band earlier this year, we held off doing so for some greater clarity on a key issue regarding compatibility between broadband and passive scientific satellite services near the lower edge of the 26 GHz band.
The delay in resolving this issue dictated a choice between, on the one hand, pressing ahead with releasing planning options early with a potentially cut-down version of the 26 GHz band, keeping Australia at the vanguard millimetre wave 5G deploying countries, but reducing the amount of spectrum available, or, on the other hand, to hold off further until we, and Australian spectrum users, have a better idea of what the harmonised frequency arrangements would ultimately look like elsewhere in the world.
In our draft FYSO we proposed to wait—at least for a while. Quite simply, the risk of releasing a sub-optimal band based on incomplete information, with the potential need to allocate the rest of the band at a later date and the resultant fragmentation of licensing that could result, was too great. We had an opportunity to avoid the mistakes of the past—to do it once and do it right—and so decided to sacrifice very short-term bragging rights for long-term efficiency and productivity.
I am pleased to say Europe has recently made some key decisions on the issue, which will allow us to make an informed decision on what to do next, knowing where at least one major economic block is headed—an essential consideration for an economy such as Australia’s. I am expecting you will see the ACMA put forward a range of options, and some challenging questions, later this quarter.
As part of our planning for the 26 GHz band, we are looking at the possible suite of licensing options that can be used to support not only a range of different 5G use cases, but also provide flexible solutions for many different types of businesses that are wanting to access spectrum suitable for 5G. Depending on the relative timing of the implementation of the new spectrum legislation and the release of the 26 GHz band to market, we may be able to offer more flexible licensing options than under the current legislation.
We are also keeping a close eye on the 28 GHz band (that is, from 27.5 to 29.5 GHz), noting it is the focus of attention in the US, Korea and Japan. The ACMA intends to progress its investigation into the suitability of some of the 28 GHz band for use by apparatus-licensed fixed wireless broadband outside of metropolitan areas. Consultation on this issue is expected to commence in the second half of 2018. This is one of the measures we are looking at to support fixed wireless broadband uses, arising from the outcomes of the 3.6 GHz band review.
Higher in the spectrum, the ACMA has flagged the possibility of changes to class-licensing arrangements to support additional uses in the 60 GHz range. This part of the band is also being looked at for 5G use around the world. These include possible arrangements similar to those of the Federal Communications Commission in the USA for the frequency range 64–71 GHz, to provide further options for data communications intended to support future 5G services. I also note that Ofcom recently announced new unlicensed access arrangements for spectrum in the 57–71 GHz bands.
We are continuing to look closely at these international developments for their application here. Given the potential amount of spectrum available in the millimetre wave bands, there is a real opportunity for Australia to look at supporting a range of different spectrum access arrangements, that address the needs of mobile, fixed wireless, satellite and IoT needs, whether on a licence or licence-exempt basis. Many exciting possibilities to consider.
Next steps
You can look forward to a number of important announcements on our spectrum planning activities in coming months, including a consultation paper on the 26 GHz band, and the finalised 2018–22 FYSO.
Like everyone here I’m sure, we are looking forward to the second exposure draft of the Radiocommunications Bill 2018. We will be publishing some explanatory material of our own around the same time to assist you in providing feedback on the draft provisions to the Department of Communications and the Arts.
And please look out for our RadComms conference, being held at the Maritime Museum in Sydney at from 30–31 October this year. This is an exciting time for spectrum management in Australia, and RadComms is shaping up to be an unmissable opportunity to work with us, and the Department of Communications and the Arts, on the future of spectrum management.
Thank you very much for your time.