A political deathmatch using copper and fibre

Two competing visions of Australia's future are being played out through the battle to rationalise either the Coalition's NBN (might we call that the CNBN?) and Labor's NBN. Ultimately, voters will land on one side or the other for largely irrational reasons.

That's not to say we are all just trusting our hearts. But voters are being asked to decide between two untested, and untestable propositions.

Turnbull's is this: copper is likely to be able to deliver everything Australians need because recent history tells us that copper speeds have increased in leaps and bounds through technological developments we never dreamed possible (mainly around compression technologies, vectoring of signals, and changes to signal modulation characteristics that 'use more' of the copper's physical bandwidth capabilities).

Conroy's is this: copper is unlikely to be able to deliver everything Australians need because recent history tells us that as fast as copper data speeds have increased (mobile speeds too, for that matter), data flows have increased faster.

No reliable cost-benefit analysis can be done to test Turnbull's vision because by definition we can't image how the apparent theoretical speed limits of copper might rise.

No reliable cost-benefit analysis can be done to test Conroy's vision because the applications that break the copper speed limits mostly haven't been invented yet (outside of certain high-end business, scientific and medical applications which are only used across bespoke fibre networks).

So toss a coin, dear reader. Your correspondent did so in 2010, and wrote several anti-NBN pieces before Doan Hoang, professor of computer networks at UTS and director of the Advanced Research in Networking Group turned that coin over for me (Time to untangle the NBN, December 2010).

Hoang explained, in 2010, that raw upload and download speeds mattered less in the NBN equation than absolute reliability, consistent speed and the lack of congestion that super high speeds help provide.

Hoang, who I put the leaked Turnbull proposition to yesterday, was once again at pains to say that he did not take a political position, but described the current all-fibre NBN proposal as a "once in a lifetime chance to create business opportunities", adding that "I hope that whoever wins the election does the right thing for the country".

To understand what he meant by that I spoke to an interactive designer, and user-experience consultant – who did not wish to be named for commercial reasons – creating front-end systems for a major telco and big-four bank.

He explained how large corporations, with literally millions of customers, have to tread carefully in the current ADSL-based internet environment.

Firstly, his team cannot design high-bandwidth applications, because so few customers would experience them the way they are supposed to be transmitted.

Nor can they design user experiences for the slowest speeds. In fact, he said, they designed a user experience for a middle ground of connection speeds, and then tested it across a range of speeds to make sure it didn't lock out the slower connections altogether.

Significantly, he pointed out that people/companies paying for higher speeds don't get anything extra from the relatively unambitious online ‘environments’ served up – they simply work the way they were supposed to, every time.

Under the Turnbull plan, released today, companies will recalibrate what they can serve up to customers. Tony Abbott and Malcolm Turnbull promised a ubiquitous 25Mbps download speed as a minimum by the end of the first term of an Abbott government.

So interactive designers wishing to use, say, 3D high-definition images of products or live, two-way high-definition interaction with a sales consultant (or more to the point, medical consultant) will make sure they don't stretch the upload/download speeds.

Following the launch of his policy, Turnbull told a special briefing for technology journalists that existing copper lines would be tested to ensure they can carry download speeds of 25Mbps and replaced if not, from within his “conservative” budget for the scheme.

Turnbull was not specific about upload speeds but said they would be around a quarter of the download speeds. That makes it a fairly ubiquitous 25Mbps/6Mbps by the end of the first term of an Abbott government rising to "between 50 and 100 Mbps by the end of 2019 in 90 per cent of the fixed line footprint".

Assuming those speeds are delivered, even to 90 per cent of households, interactive designers trying to optimise user experience, and thereby revenue generated from each encounter with a customer, will have a new 'middle ground' bandwidth to limit their imaginations.

At the launch today, Abbott made much of fact that even the 25Mbps download speed would be enough for four channels of HD TV, delivered to four different screens in a house at about 6Mbps (download) per video stream.

Well, yes and no. It will as long as there is no variability in the speeds obtainable from a copper loop running to a house (for example, when it rains), and that nothing changes with congestion on the backhaul network or through 'pair gain' setups, in which signals on one copper line are split into numerous lines (the latter is something Turnbull promised would be fixed where required).

Conroy's argument sends a different message to the online/interactive designers that create the interfaces between online businesses and their customers. Conroy's all-fibre solution, that can deliver 100/40Mbps to premises connected, and up to 10 times that speed in future, effectively says ‘design whatever you like and 93 per cent of customers will have a great experience’.

Turnbull's plan, by contrast, effectively says 'you won't need more than these speeds, and can always upgrade if you do'. The cost of the upgrade is debatable. In the UK, BT is offering a similar upgrade service, costing up to £3500. Converted at today's exchange rate that would mean a local cost per dwelling to upgrade from copper to fibre of up to $5100.

However, the dramatic slump in the British pound in the past few years favours the Coalition in this conversion. Using full-time earnings as a proxy for the cost of a team of workers to install fibre (a rough purchasing power parity conversion based on current full time wages of £26,500 and $72,592) would suggest the cost of a connection could be much more – up to $9600.

The other side of the ‘connect fibre if you need it’ proposition is that fibre-to-the-node is much, much cheaper to build. The Coalition estimates its plan will cost only $29.5 billion (nominal dollars over the build time), and that it will complete the build by 2019 – two years ahead of Conroy's planned finish line of 2021.

The Coalition has also projected Conroy's plan to cost $94 billion in nominal dollars, based on what it sees as four “fair” assumptions:

– Users won't want the faster speed packages on the NBN so will pay less to connect, thereby creating revenue shortfalls for the Conroy NBN.

– It thinks connections in built-up areas will turn out to be much higher than estimated by the NBN's corporate plan ($3600 per dwelling).

– It thinks around twice as many homes will choose to use only wireless devices (3G, 4G and soon LTE-Advanced) for all their communications, eating into NBN revenue.

– And it thinks the NBN build will go four years over schedule.

In answer to these assumptions, Conroy insists the targets in the corporate plan will be met – that is, not four years late, not with lots of people refusing to connect, not at installation costs higher than in the plan and not with subscribers choosing only the base plans.

The present time overruns were explained by Mike Quigley when he announced the new rollout figures last month as being due to a "bottleneck" with contractors failing to employ "fibre splicers" to keep up with their own build commitments. NBN Co has since employed an 80-strong "tiger team" of splicers to fix the problem that, Quigley says, the contractors were failing to.

We will know whether or not that has accelerated the rollout "within a month", according to Conroy's office, when new figures for premises passed are released.

Until then, the Coalition's list of assumptions, and Conroy's rebuttals of them, are difficult for even confirmed NBN geeks to decide between.

As things stand, voters will accept Turnbull's 'that's fast enough' proposition or Conroy's 'no it isn't, which is why you need to pay more and wait longer', based on their own vision for what's achievable down a piece of copper or fibre.

And in the background will be the daily spectacle of Conroy and Turnbull slogging it out for the hearts, minds and imagination of the nation.