A retort to Bill Morrow’s Gigabit NBN “facts”

NBN CEO Bill Morrow made a flippant remark about gigabit internet recently, saying that “even if we offered it for free, we see the evidence around the world that they wouldn’t use it anyway”. Unsurprisingly, this caused a massive uproar in the tech press and the day after, Morrow posted on the NBN blog explaining why he said that.

No surprises, but the post is full of typical NBN spin. A text book example of an organisation cherry picking facts to suit its argument. I filled in some of the gaps Morrow conveniently left for us his blog post.

There has been a lot of discussion in recent days around the need for Gigabit broadband in Australia – let’s set some things straight…

Translation: I said something stupid and now I have to explain myself.

The nbn™ network is here to bring fast broadband to all of Australia by 2020 – more than 4 million premises already have access to fast broadband via the nbn™ network today. In just a few months we will be at the halfway point of the rollout.

NBN was supposed to be bringing fast broadband to all of Australia by 2016, wasn’t it?

Next year, we will be three quarters built. By 2020 we will be complete. What once seemed impossible is now a reality.

Impossible? Nobody ever said it was impossible. Who said it was impossible? People said it would suck (turns out they were right), but never that it couldn’t be done.

I was asked last week by the media about the need for Gigabit (Gbps) speeds in Australia. These are lines that are 40 times faster than plans based on our most popular 25Mbps wholesale service. The fact is, nbn already offers a wholesale 1Gbps product to retail service providers, which RSPs can make available to more than 1.5 million homes, and has been on sale for around four years.

This is true. 1000/400, 500/200 and 250/100 speed tiers have been available on the fibre technology for a while now. However, it is disingenuous to say 1.5 million homes can have access to gigabit speeds and that it’s been on sale for around four years. There’s only 1,117,000 (as of December 2016) homes that have a fibre NBN lead-in. Have ~400,000 homes received fibre in the last 2 months?

Currently, there are no retail 1Gbps speed plans on offer from the retailers. This is, in our opinion, because there is still minimal consumer demand for these ultra-fast speeds – especially at the prices retailers would have to charge for them.

The prices retailers have to charge for them is set predominately by you! There’s a reason why all the NBN plans cost roughly the same from all the ISPs. If you’re interested to know the costs for an ISP to connect to the NBN, one of the few things the NBN still releases publicly is a wholesale broadband agreement price list.

The reasons why ISPs haven’t offered speeds greater than 100/40 are complicated. This article on PC & Tech Authority explains it well. Simply put, it costs too much for the ISPs, as there’s not enough areas with gigabit capable connections to make it worthwhile, combined with the excessive cost that doesn’t match what it actually costs NBN to operate.

*Our own data shows that 83 per cent of people on services powered by the nbn™ network today are ordering retail services based on the two lowest wholesale speed tiers 25/5Mbps and 12/1Mbps. There is currently around a $20-$30 monthly difference between what retailers charge for plans based on our 25/5Mbps wholesale service and our 100/40Mbps wholesale service – but only 13 percent of end-users are willing to pay that relatively small premium for access to that faster tier. *

Could this be because most people can’t actually get 100/40 speeds? Those on satellite can’t get anything more than 25/5 and wireless tops out at 50/20 (both of which are horribly congested by the way and real world speeds don’t match the advertised speeds).

100/40 on FTTN, well, how long is a piece of string? While theoretically capable of up to 100/40, the reality is that speeds are much lower. Someone may decide to order a 100/40 FTTN service, but when it’s made active, will see that their VDSL sync speed is only 18/2 or 30/4. so they downgrade to 12/1 or 25/5 – why pay $30/m extra for speeds physically impossible?

We have to be realistic about what the market is actually telling us about demand for ultra-fast services. Of course, the demand for Gigabit speeds could, and probably will, change – current plans for FTTP and DOCSIS 3.1 HFC suggest they will be able to deliver such speeds to around 5 million premises on the nbn™ network by 2020 and the other parts of our network, with the exception of satellite, have upgrade paths to offer the same ultra-fast speeds when demand comes around.

Ahh the good ol’ upgrade path promise. What’s the upgrade path for FTTN to reach gigabit? G.Fast? A technology just deployed in Switzerland which when installed in the real world never actually hit gigabit speeds? Which will still have the same issue as VDSL where unless the conditions are pristine, you won’t see anywhere near gigabit sync speeds? Plus it’ll require moving the node ever closer to the home, at which point, you may as well have run fibre once you added up the cost of installing the original VDSL node and the new G.Fast node!

Rather than build for a demand that may materialise in 10 years, we are constructing a national network capable of continuous upgrading to meet market needs as and when they arise. There is little point in adding to the already high $49 billion cost of the nbn™ network to provide a capability that end users do not yet require and RSPs are not selling.

You say there’s little point in spending more than the already high $49b to provide capability users don’t need, but what’s more irresponsible is spending more money later, when it can be done properly now. While I was writing this, the much more clued in than me on this topic, David Cooper, worked out that it would cost $12-15b to upgrade FTTN to FTTC (aka FTTdp) for G.Fast.

Would it not be cheaper just to roll out fibre to those premises now and save the hassle and expense of going back in 10 years? Plus the bonus of those users actually enjoying gigabit speeds now rather than in a decade? That also assumes G.Fast would actually reach gigabit speeds…

We spend a lot of time researching overseas markets about Gigabit speeds, and the trends are not what you might think. We have met with global operators offering 1Gbps services and they willingly concede that their end-users are simply not using speeds anywhere near 1Gbps – in fact, they would be lucky to actually ever use a fraction of that speed – so, at present, the Gigabit game is really about marketing, not actual utility.

So you’re saying that the ISPs who have invested in gigabit fibre in countries like: Hong Kong, Japan, Singapore, South Korea, the Czech Republic, Estonia, Finland, France, Iceland, Ireland, Latvia, Lithuania, Norway, Portugal, Romania, Serbia, Slovakia, Spain, Sweden, Switzerland, Turkey, the UK, Israel, Canada, the USA and New Zealand are just doing it for marketing purposes? Really? They spent all that money just becuse it looks good in a brochure or a powerpoint slide? Get your hand off it mate.

Even in a heavy usage household, right now it’s likely you’d struggle to generate the need for anything close to 1Gbps – if you had five 4K TVs streaming 4K movies simultaneously then that’s only around 100Mbps being consumed – leaving 900Mbps idle. Given that the vast majority of current online video viewing is in SD or HD – requiring only 2Mbps-5Mbps then a 1Gbps pipe would be enough to stream 200 HD streams simultaneously – way, way beyond the requirements of a normal household.

Yes, because all we need 100mbit+ speeds for is watching 4K video. There’s no other need for gigabit besides watching a few Netflix vids aye?

Don’t tell the engineers and architects transferring huge detailed drawings. Don’t tell the video effects artists sending massive video files between clients overseas. Don’t tell the educators wanting to export Australia’s world class education content across the world. Don’t tell the other professionals wanting to host meetings with overseas clients. Small businesses wouldn’t use gigabit speeds at all, would they? God forbid there’d be multiple people in the same premises wanting to heavily use the internet at the same time!

Then there’s the things we haven’t even thought of yet. Pew Research explores the unknown future uses far better than I could – which really, is what the NBN is all about. A platform for innovation. Why knee-cap it now, only to come back in 10 years and spend more more on it then, when by your own admission know the need for it will be inevitable?

Of course, if we were in a position to deliver 1Gbps for $49/month, as they do in Singapore, then we would do it – but we are simply not in that position from an economic point of view. The fact is that Singapore is a city state, comprised of high density apartment buildings – and Australia is over 10,000 times larger than Singapore. The Singaporean Next Generation National Broadband Network [NGNBN] cost only a fraction of what it will cost to build the nbn™ network.

Population density in the inner ring of our capital cities is actuallly pretty high.

Mosman, Lane Cove, Burwood, Strathfield, Woollahra, North Sydney, Waverley, Willoughby, Canada Bay, Randwick, Georges River, Bayside, Inner West Council, Cumberland and Canterbury-Bankstown in Sydney. Port Phillip, Yarra, Stonnington, Glen Eira, Melbourne, Moreland, Boroondara, Darebin, Bayside, Moonee Valley in Melbourne. Subiaco and Vincent (with Perth, Fremantle and Mosman not far behind) in WA. Holdfast Bay and Unley in SA.

All these places have over 2800 people living in a square kilometer. In Melbourne and Sydney, density is growing rapidly too. More than enough density to give every single premise gigabit fibre. Sure, it’s not Hong Kong or Singapore levels, but they are roughly the same as most of the other locations that do have gigabit fibre.

The nbn™ network is costing around $49 billion to build – and we need to recoup that cost – given that our business model is split between driving revenues from access and consumption charges, we simply cannot match the kind of 1Gbps pricing on offer in markets like Singapore and Hong Kong.

That business model is imposed by the government. There’s no inherent need for the NBN to turn a profit. We don’t run hospitals at a profit, or schools, or our defence force – but they all add to Australia’s prosperity. That’s not to say the NBN shouldn’t be somewhat self sufficient (i.e: maintenance and support), but the argument that it has to turn a profit is not the default argument and can be changed.

In Hong Kong, you can deliver FTTP for as low as $150 per premises. In comparison, thanks to Australia’s unique geographical and population density situation, it is costing us an average of $4,400 to connect every FTTP premises and can cost significantly beyond that in many situations – these are the stone cold facts of the matter and there is no getting away from them.

The cost of installing fibre in greenfields isn’t that high. The cost of installing fibre was actually dropping. In NZ they managed to get the cost of a fibre install down from ~$4700 to ~$3000. Progresses in technology such as skinny fibre and improved connection methods reduce costs even further. Even the NBN’s own trials have shown savings in the cost of FTTP installs. Mike Quigley’s post-mortem of the NBN in June 2016 outlines the progresses made in reducing the cost of fibre installs. This information was of course leaked, as we can’t have anything offical saying fibre is great can we Bill? That’s not the narrative NBN wants out there.

If you want evidence that we are taking the right approach, you need only look at what is happening in the US with Google Fiber. Last October, the internet giant announced that it would be putting its much publicised FTTP deployment on hold and would look at new ways to deliver ultra-fast broadband. Google Fiber found that viable economics for Gigabit broadband simply did not exist once the initial hype of the project had faded and the product hit the streets.

It’s not because of the cost, but because of the entire project was a side show and not part of Google’s reason to exist (ads are why Google exists). The regulatory environment in the USA cut the legs off Google fibre and they had no desire to turn into a government lobbying machine like Comcast, Time Warner and Verizon and AT&T.

The NBN however is a government company, large swathes of law was changed so NBN could exist and prosper. In fact, NBN just a few months ago had a levy passed so faster, more progressive telcos can’t compete and eat into NBN’s bottom line. Google, for all the money and smart people they possess, can’t do that.

Indeed, respected writer Farhad Manjoo pointed this out in an article published through Slate in 2013. “[Google Fiber] felt a little underwhelming. After all, who needs to play five HD videos at the same time? If that’s Google’s best demo of its superfast service, what does it suggest about what regular people will do with it? What’s more, the demo didn’t even begin to approach the limits of Google Fiber—with five HD videos playing simultaneously there were still hundreds of megabits left on the pipe. When I got back home a few days later, I replicated the same test on my home broadband line and experienced only a few hiccups.”

The best story Bill could find to back up how unnecessary gigabit was an article from 4 years ago. Look at all these articles from these groups who have gigabit and love it or are so desperate for it, build it themselves:

So while a journalist at home might not find gigabit internet all that amazing, there’s piles of evidence that the money spent on gigabit level internet brings vast economic improvements.

Ultimately even a company with a huge amount of capital available, such as Google Fiber, found that digging up driveways and gardens to connect every single premise to FTTP was too expensive and time consuming.

Google ultimately concluded that they could not make it work commercially offering Gigabit FTTP broadband at just $70/month. We are still waiting to see what Google Fiber will do next – reports suggest that it will move towards a Fixed-Wireless solution – but quite clearly their conclusion was that the economics of delivering Gigabit speeds via FTTP simply did not add up.

Read what I said above about how it wasn’t the cost impeding Google Fiber, but incumbent telcos and corrupt local governments. Besides, if fibre is so bad, why didn’t Google deliver gigabit over FTTN or HFC? Oh that’s right – they can’t because the law doesn’t let them. That’s literally the reason why they want to use wireless – not because it’s better (it’s not), but because it lets them bypass all the 20th century archane laws designed to protect rent-seekers.

That’s not to say that Google Fiber didn’t have an impact in the market – their presence forced the incumbents to raise their game with the likes of AT&T now marketing FTTP access to 4 million premises across 46 metros, but this still represents only 3 per cent of the near 130 million premises in the US – we are talking about a niche market.

A niche market! I thought nobody wanted it? Now we’re on to niches, cool. The US FCC knows that gigabit internet is important and it even admits that incumbent meddling is a big problem in the USA. The new chairman of the FCC knows this too and wants to create “gigabit opportunity zones“, where ISPs are encouraged to build gigabit networks and laws are changed to reflect this.

Markets like Hong Kong, South Korea and Japan started their deployment of high-speed broadband in the late 1990s and having fully recouped their investments in delivering FTTB with VDSL – typically offering up to 100Mbps – are now delivering 1Gbps and even 10Gbps via FTTP.

Mmm, yes, funny that. Our smarter Asian neighbours realised how great the internet is for boosting their economies way before we did.

For a variety of reasons, our broadband upgrade in Australia started much later, so we cannot judge ourselves against markets like these; they are much further along on their journey and you just can’t compare Australia to Singapore or Hong Kong for obvious reasons including those stated above.

You do realise that we compete with these countries economically? It’s fine to hand-wave away some technical limitations and historical blunders, but they are excuses. We have the opportunity to catch up and you’re pissing it away because it’s too hard.

This is why operators around the world are so excited about technologies like DOCSIS 3.1 and G.fast, which allow Gigabit broadband to be deployed at substantially lower price points and in far less time than it takes to deploy Gigabit services over FTTP – and we will see both of these technologies emerge much more fully in the next few years globally as well as here in Australia.

So gigabit is great, as long as it’s not over fibre? I thought the argument is about gigabit’s viability, not the medium under which it is accessed.

We know that people in Australia want access to fast internet – that is precisely what we are aiming to deliver with the nbn™ network. We also know that 1Gbps speeds are simply way beyond what even the most advanced end-user needs today, let alone what is needed by regular families across Australia.

Regular families, aww. Yes, regular families are the reason we spend $50b on infrastructure. The NBN isn’t about regular familes you dipshit, it’s about making Australia’s economy compete in a global market. Who cares if Garry and Jane and their idiot kids can access Netflix? That’s totally not why the NBN exists and to disregard all the other people shouting for faster internet that aren’t “regular families” is an insult to their economic contributions to this country.

There is literally not a single mass market consumer application – or even a combination of applications – that requires 1Gbps capability right now. This will change in the future, especially as augmented reality, artificial intelligence, 8K technology and virtual reality penetrate our daily lives. In fact, we hope that by delivering a fully connected continent we can help create a market for Gigabit applications – and when Australians need those kinds of speeds we will have the solutions in place to provide it.

So.. you’re creating a market for gigabit speeds, yet, don’t think it’s worth investing in that right now. At least you changed your mind from “may eventuate” earlier on in your rant, to “this will change in the future”. Let someone else worry about it when it’ll cost more and take longer. That’s exactly how we should run all our national infrastructure projects. Build the bare minimum for now and then spend way more money later when it’s someone else’s problem. Can’t wait for it.

Thanks for fucking up my generation’s economic future Bill Morrow and everyone at the NBN. I’ll remember you when there’s massive levels of unemployment and high taxes because all the shiny rocks in the ground are no longer needed, the property bubble burst and the remaining profitable industries pissed off to NZ and Singapore because they had the foresight to invest in proper broadband and we didn’t.