CommsDay Summit 2015 – Keynote address
20 April 2015
It's a pleasure to be here, and I want to thank Grahame Lynch and CommsDay for inviting me to this conference - always a highpoint in the industry calendar.
This is my eighth speech to a CommsDay event as shadow minister or Minister for Communications.
In my address last year, I said I was increasingly – although somewhat cautiously – confident about the future of the National Broadband Network.
Since then, the Vertigan Panel has released its Cost-Benefit Analysis of Broadband, showing that a multi-technology mix rollout is the most future-proof strategy for delivering universal high-speed broadband and has net economic and social benefits of about $18 billion dollars.
The Vertigan Panel's Market and Regulatory Report has laid out a roadmap for a more competitive future broadband market.
And in December, NBN Co signed landmark agreements with Telstra and Optus that facilitate the shift to a multi-technology mix or MTM NBN. The deal allows NBN Co to gradually take ownership of many parts of Telstra's copper and HFC cable networks.
So we've laid the groundwork and in partnership with NBN Co's Board and management are turning this project around – not an easy task given the state it was in at the change of government. There is still a lot of work to do, and a good deal of risk associated with the project which I'll expand on later.
But first let me report that thanks to our more flexible and pragmatic approach to the NBN, we're seeing real progress on the ground.
2015 is a big year for the rollout, which is gradually accelerating:
- There are now 913,000 homes and businesses across Australia which can get the NBN. That is almost four times the number in September 2013.
- As of the 9th of April there were 394,000 active customers using the network – again, a four-fold increase under this Government.
- NBN Co recently added an additional 550,000 premises to the construction schedule.
The Vertigan panel's Cost-Benefit Analysis indicated that an NBN leveraging a mix of different access technologies was the right model for Australia. Practical experience is confirming that it increases the speed and decreases the cost of the rollout.
Upgrade Paths Around the World
Elsewhere, however, carriers and vendors are not waiting on economic modelling to dictate technology choices. Competition and customer demands are driving decisions. And pragmatism is the order of the day.
Certainly there is no evidence of the fervor for fibre to the premises our critics sometimes claim to see in overseas markets. The European Fibre to the Home Council, in an update to its 2018 forecasts last May, acknowledged that although there have been upward revisions in forecasts for Latvia and Spain, overall the outlook for FTTP in Europe is down.
In particular, the report cited Germany, the UK and the Netherlands as key markets which had seen substantial downward revisions. In all of those countries – which have existing network architectures similar to Australia – deployment of FTTP at scale has turned out to be much more time consuming, costly and disruptive than originally expected.
For carriers, the inability to upgrade customers quickly and cost effectively is existential, as cable operators eye upgrades to DOCSIS 3.1 available at relatively low cost with minimal civil work and customer disruption.
Rather than 'big bang' projects there is a renewed interest in incremental upgrades where legacy infrastructure can be leveraged.
On this front the most exciting emerging technology is g.fast, which as most of you know, uses a wider chunk of copper's available spectrum – typically the 106 MHz profile as compared to the 17a profile for VDSL – meaning users can achieve much higher bitrates, albeit with greater attenuation.
With the g.fast standard now settled and approved by the ITU, there is growing interest around the world.
The recent expansion of the UFB project in New Zealand from 75 per cent to 80 per cent of premises included a 'registration of interest' document which indicates g.fast is among the technologies being considered.
Elsewhere BT Openreach has two trials of g.fast planned for later this year. And Swisscom recently informed us it had achieved download data rates of 382 megabits per second and upload rates of 112 megabits in a g.fast trial in Riggisberg. Swisscom aims to achieve cumulative capacity (the sum of download and upload bandwidth) of 500 megabits per second over 220 metre copper loops using a fibre-to-the-street topology.
These are very promising developments which we are following closely. NBN Co's rollout is being planned to allow for fibre-to-the-distribution point and fibre-to-the-basement, both of which can in due course leverage the g.fast protocol.
At the moment, there has been no widespread commercial deployment of g.fast, but the level of interest across the industry indicates we won't have long to wait before there is.
Progress with new access technologies
NBN Co's shift to the multi-technology mix rollout is well underway. Progress is being made integrating, upgrading and from early 2016 utilizing the Optus and Telstra Hybrid Fibre Coaxial networks that pass almost 4 million premises in Australia's six largest urban centres to deliver the NBN. Investment in these networks will enable very high-speed broadband using DOCSIS 3.0, at lower cost than the access technologies currently used in the NBN.
I'm particularly excited about NBN Co's recent announcement that in 2017 it will upgrade its HFC networks to DOCSIS 3.1 (a world first) and be able to offer Gigabit speeds to those users who demand them.
Today's standard data protocol on cable networks, DOCSIS 3.0, offers high bandwidth, but 3.1 is even better. Indeed, the attractive upgrade roadmap HFC offers over coming years confirms the sense in obtaining and cost-effectively upgrading this excellent existing infrastructure that under the previous government's policy was to be scrapped or prevented from carrying broadband.
NBN Co recently announced HFC construction trials in four suburbs across Queensland and New South Wales, which commence shortly.
Elsewhere, early fibre to the basement services for multi-dwelling units have achieved average speeds of 89 Mbps download and 36 Mbps upload.
NBN Co has now officially launched its much-heralded FTTB product, which is available to about 2,000 premises, and more buildings are being added rapidly.
The level of service demonstrated over FTTB would allow a household to stream around 18 high definition television streams concurrently.
Markets such as Japan and South Korea have long embraced FTTB, and around a third of Australian residences are in multi-dwelling units, so its local rollout is overdue. Again, this leverages existing infrastructure to provide fast broadband sooner and at less expense to millions of users.
NBN Co's 2014-17 Corporate Plan indicated 1.4 million premises around Australia would eventually be connected via fibre-to-the-basement or fibre-to-the-distribution point.
NBN Co is also getting on with the job in rural and regional areas, and recently announced it will be upgrading its fixed wireless product offering. Currently there are 194,000 premises covered and 35,000 premises connected to this network, with customers indicating high levels of satisfaction with the service. The fixed wireless rollout will be about 50 per cent complete by June.
Starting next month, fixed wireless customers will be able to try NBN Co's improved service offering download data rates of 50 megabits per second and upload rates of 20 megabits per second. This pilot will be available to all consumers on the current 25 megabits per second tier whose retail service providers choose to participate. Neither consumers nor NBN Co will need to spend anything.
The pilot will take place from May and the commercial launch is currently planned for late 2015.
Higher bandwidth for people living in regional and rural areas is a priority for the Company and the Government, and provides them with new opportunities to participate in the digital economy.
Limitations of 'Last Mile' Upgrades: The Netflix Conundrum
Higher bandwidth over the access network is of great interest to Australians with substandard last mile connections and the engineers in the room. What customers as a whole care most about, however, is the quality of their experience and of the services they receive.
That is why the launch of Netflix in Australia is of interest to everyone in the broadband ecosystem, be they carriers, vendors or content providers.
According to figures compiled by Sandvine, as of last week Netflix accounted for a third of all Internet traffic in the U.S. On the first Sunday night following launch of the new season of House of Cards one operator saw a 30-35 per cent spike in traffic. Meanwhile, Cisco's latest VNI projects that by 2018 streaming video will account for up to 91 per cent of consumer IP traffic.
We can see the same types of trends coming in Australia. Already 30 per cent of Australian homes have 'smart' – or internet capable – TVs, an increase of 7 percentage points from last year. And more than 13 million Australians are now watching video on the internet, consuming an average of almost 7.5 hours each month. It is forecast that by 2023, the average Australian will watch at least a quarter of their TV online.
According to iiNet, Netflix already accounts for 15 per cent of its traffic .
But again, the link between the volume of data downloaded in Australia and line speeds required is generally very poorly understood and many people still mistakenly think that as data volumes increase last mile line speeds will need to increase in a linear fashion. They don't and they won't.
The launch of Netflix once again provided a soapbox for those who claim only fibre to the premises will unlock the benefits of the digital age. What a pity the facts don't support their contention.
Mark Pesce, for instance, writing in the Register, claimed that an ultra HD stream would require 15-25 megabits per second download capacity and stated:
"Even the 50 megabits per second promised as the cheap-and-cheerful alternative to Australia's fibre-to-the-home National Broadband Network won't accommodate two UHD streams into a home simultaneously - never mind the kinds of stresses they'll put on a neighbourhood node when every home in the suburb tunes into two UHD content streams every night of the week." 
There are a couple of problems with this. First, a 50 megabits per second download connection would quite clearly permit two simultaneous UHD streams with overhead to spare.
Secondly, Mr Pesce is confused by network topology. In an FTTN network the customer has a dedicated copper line from the node. From the node to the Point of Interconnection customers' traffic will travel over a shared fibre channel. Contention on that shared channel is possible but much less likely than the most common source of contention which is in the transit or backhaul network from the PoI.
In a world where ISPs will compete on price and service, the far more likely choke points for services such as Netflix will be in the backhaul network and the choices retailers make about CVC and backhaul contention ratios. That has certainly been Netflix's experience in the United States.
And as for Australia, in the words of Telstra's head of networks, Mike Wright, retailers like iiNet who blame the quality of the access network miss the point, given it's only during periods of peak demand that end users experience degraded service:
"iiNet has end-to-end responsibility for their network, including managing the various network elements required to deliver reliable video streaming, one of which is traffic backhaul and interconnection. It is the responsibility of each ISP to have sufficient backhaul capacity to deliver services to their customers."
One of the consequences of the NBN is going to be to demonstrate to consumers that iron-clad guarantees about performance cannot be made on ANY access technology, given the way factors such as CVC pricing and backhaul contention choices play into the end user experience. That may burst the bubble of those who claim FTTP solves such problems, but it is reality.
The importance of completing the NBN as cost-effectively and quickly as possible cannot be overstated. As I've said many times before, the current Government would not have chosen to upgrade our communications networks this way, and government owned start-ups are not the ideal (or conventional) vehicle for building telecommunications networks - but we have had to play the hand we were dealt and are doing so.
While there have been very large gains over the past 18 months in the professionalism and capabilities of NBN Co's leadership and team, we shouldn't blind ourselves to the reality that this remains a very risky project – particularly given the onerous financial obligations the Company entered into under the Labor government.
The NBN was risky in conception, it is proving to be risky in execution and it will remain risky even after it is completed, given the dynamism and rapidly changing character of the telecommunications sector. The Government is currently considering an initial draft of the NBN Co 2015-2018 Corporate Plan, which confirms the construction, commercial and technological uncertainties the company faces.
The NBN is the largest infrastructure project in Australia's history. The challenges of turning it around have been massive and there are further challenges ahead, including more legacy issues that will need to be dealt with. As we deal with them, we remain guided by our commitment to deliver affordable high speed broadband as soon as possible and at the least possible cost to taxpayers – so all Australians can access the benefits that high-speed broadband provides, including its potential to boost economic efficiency, productivity and innovation.
While there's more work to be done, the Government's practical and agile approach to the NBN is already achieving results.
 European FTTH Council, (2014), "European FTTH Forecast 2013-2018: Behind The Numbers", p.24
 In Germany, Deutsche Telekom is planning to roll out VDSL vectoring (allowing download speeds of 100 Megabits and upload speeds of 40 Megabits) to an extra 5.9m homes, raising household penetration of high speed internet from 65% to 80% by 2018.
 France also saw a minor downward revision.
 Incumbent networks in Spain, for instance, for historical reasons were point-to-point in order to avoid street distribution points, meaning their ducts are very large.
Sandvine Blog, (2015), "Sunday Night Streams: How Did HBO Now Fare?" http://www.internetphenomena.com/2015/04/sunday-night-streams-how-did-hbo-now-fare/
 Sandvine Blog, (2015), "House of Cards Season 3 Premiere",. available online here: http://www.internetphenomena.com/2015/03/house-of-cards-season-3-premiere-weekend/