- How much broadband speed do Australians need to ensure the nation’s future prosperity, in the widest sense of that word?
- How close is that future? That is, how soon will we need that speed?
- How much are willing to pay, both personally and from the collective government purse?
- What is the best way to achieve that goal, given we’re starting from a real-world position that includes a commenced-but-not-much-more fibre to the premises (FTTP) access network and a bunch of related contracts?
- Both sides are making statements about technology choices, capabilities and prices, as well as business models. How can we be certain that their proposed answers to the third and fourth questions will indeed deliver what’s needed to match our preferred answers to the first two?
The Coalition’s policy and the accompanying 36-page backgrounder present a coherent and powerful argument. Most of it will be broadly familiar to anyone who’s been paying attention to what Turnbull has been saying for the last three years?—?as opposed to just whining that he doesn’t get that FTTP is technically superior. He knows it is, but his argument is that technical superiority is not the only factor that needs to be considered.
Dispensing with the technical issues first, there’s the unarguable fact that NBN Co’s FTTP runs at 100Mbps today, and maintains that performance regardless of the customer’s distance from the exchange. It’s easily upgradeable to 10 times that, 1Gbps, by replacing the endpoints. Google Fiber, as they spell it over there, is already rolling out 1Gbps in selected US locations.
The speeds achievable with fibre to the node (FTTN) can be highly variable, depending on the quality of the final copper cable segment, distance from the node to the customer’s premises and the technology used for that final section?—?but whatever technology used, it can never be more than about 100Mbps thanks to physics.
Yesterday Simon Hackett, founder of internet service provider Internode and now a director of iiNet, included this cheeky chart in his presentation to the CommsDay summit. It plots the maximum data speeds achievable with different technologies versus distance, FTTP (labelled “FTTH” for fibre to the home) versus the various xDSL technologies that form the final part of FTTN …
It’s cheeky because with FTTN you don’t run the xDSL stuff for more than about 600 or 800 metres, so we’re really only looking at the left-hand 15th of the chart. Still, it makes the point. FTTP’s potential is all the way up the top, FTTN’s is all the way down the bottom.
Which brings us back to our core questions. Turnbull points out, quite rightly, that even an entry-level 25Mbps download speed is more than enough for current home needs and then some?—?including several simultaneous high-definition video streams, the traditional example of a capacity hog.
He acknowledges some users may demand more?—?he cites architects or software developers shunting large files around?—?but argues the needs of these few don’t justify the more expensive and slower-to-deploy infrastructure of FTTP for everyone when it’d be quicker to do a FTTN rollout and get at least some improvement sooner for more people. Anyone who needs even more, and sooner, can pay for it?—?after all, it’s a legitimate business expense.
Turnbull tends to avoid discussing the issue of upload speeds, as does Labor. Solid upload speeds are important if you’re creating and distributing large files, or streaming out your own live video, rather than passively watching stuff, but that depends on how the retail broadband products are configured and, again, there’s an argument that creators can be dealt with on a case-by-case basis rather than providing the capability for everyone up front.
Counter to this argument is the idea that more bandwidth will encourage the development of services we have yet to imagine, benefiting all. One past example is Skype: the video chat product couldn’t be developed and deployed commercially until a critical mass of users with enough bandwidth made it worthwhile. Turnbull’s backgrounder hoses that down:
“The greatest uncertainty overhanging investment in next?generation fixed networks is not over costs, despite the mystery that surrounds NBN Co’s refusal to be open about the full capex cost per premise of its rollout. It is over revenue and user demand; nobody knows what applications may emerge to use the bandwidth that fibre provides, or what value consumers might place on them.
“At the moment the only answer to this quest for a ‘killer app’ for fibre is more and better television.
“In markets such as the US, where cable TV subscriptions are an embedded part of consumer behaviour, this may tip the scales in favour of investment. But relying on TV in addition to broadband and voice to drive revenues is less persuasive in Australia, where subscription TV penetration is lower.
“Uncertainty over demand is why many private investors are choosing to defer investment in FTTP and, if they are incumbents or otherwise able, to instead deploy FTTN. FTTN puts less capital at risk. Deferring an expensive and irreversible investment commitment such as FTTP, to the extent this is possible, is prudent commercial management if it permits more information about actual usage and investment returns on fibre to emerge.”
In brief, Turnbull would deride this as “build it and they would come” wishful thinking. Proponents might well counter “built is and it becomes possible”. And Turnbull’s rejoinder would be all about fiscal prudence.
Prudence is the core of the Coalition policy, and about its comparison with Labor’s NBN. The backgrounder steps through the various scenarios and spells out the assumptions. Those wishing to claim Turnbull is wrong need to step through those and point to the errors?—?and Turnbull’s approach is efficiently business-like. Turnbull told the tech press briefing yesterday:
“Everything’s obviously very sensitive to revenue, right? So we’ve assumed the same revenue assumptions for both Labor’s NBN and for ours.”
He’s also assumed a global interest rate of 4.5%. They’re assuming it’ll cost $900 per premises to run FTTN past it. Turnbull again:
“That is a very conservative assumption. That is a lot more than is being spent in northern Europe, or the United States, to pass a premise with FTTN.”
For Labor’s NBN, they’ve assumed $3600 per premises for FTTP.
“I think that is pretty generous to them, higher than in their business plan but, I have to say, whether you look at their experience to date, where you look at what Testra’s done in South Brisbane, where you look at what’s happening in New Zealand, you could quite credibly put in a materially higher figure than that.”
But the model is highly sensitive to per-premises prices. It’s when you plug in $4000 or even up to $5000, and add in other adverse options, that you start to see the blowouts to $90 billion The Daily Telegraph has been reporting. Turnbull has simply described the potential outcomes and noted the risks; the Tele did the pick-and-screech.
Hackett summarised the key objections to the Coalition plan. FTTN isn’t as simple as its made out, with problems ranging from it having double the power consumption of FTTP, to having to negotiate the siting of 60,000-odd node cabinets with local councils, to having to service the nodes’ batteries and more.
And then the whole FTTN shebang would have to be upgraded to FTTP anyway at some point?—?unless demand ”magically plateaus”, as Hackett puts it, at around 2x to 3x the current level. History has shown a continuous, unbroken increase. Any new data infrastructure that’s deployed has its capacity filled sooner rather than later. And yet assuming that will continue to be the case is a belief, not science.
Turnbull has factored in operational costs sourced from existing operators. He’s also provided, on page 15, his working-out to challenge the notion that building FTTP in one hit now is, overall, cheaper than a two-step approach?—?using the kind of modelling that investors are used to seeing. It’s there to be challenged.
Ultimately, it’s a choice between two sets of beliefs. The Coalition’s mix naturally includes “private ownership and an open, competitive market create more direct incentives for efficiency and lower prices for consumers” and “private investors are usually better placed than taxpayers to evaluate, fund and bear the risk of new investments in networks or technologies in an increasingly dynamic sector, and they’ve presented a business plan. Labor’s naturally includes issues of equity and building infrastructure for the community good. That’s been matched with the not-unreasonable technologists’ beliefs about the empowering and enabling nature of data networks?—?but they are nevertheless “just” beliefs.
And in any event, how much of this stuff should the government pay for, versus individual householders and businesses?
Between now and 14 September the broadband battle will be fought on these two fronts: left-brain hard numbers; and right-brain beliefs, gut feelings about the benefits versus the risks, and ideology.
This article first appeared on Crikey.