Conroy's fading NBN vision

When Kevin Rudd and Stephen Conroy came up with the extravagant Telstra-smashing plan for a $43 billion fibre-to-the-home broadband network in 2009 they knew the NBN would stamp their names on history.

A year on, however, Rudd's name has made history in an unforeseen way, and while the Communications Minister battles on with the plan under a new PM, there are signs that even he might not get the nation-building historical reference he wanted.

In a joint-party-room meeting yesterday, the Coalition finally took the plunge and adopted a policy that would split Telstra in two and pave the way for a 'fast' broadband network considerably cheaper than Conroy's $2000 per capita plan.

This will give Malcolm Turnbull the formal backing he needs to provide a credible, less-costly and fast-enough broadband policy that will genuinely hold the government to account on this issue.

The Coalition policy is based around hiving off a wholesale arm of Telstra – dubbed 'CanCo' for the time being, according to The Age – to provide all Australians with broadband services at a mandated speed of 12 Mbps.

That's not terribly fast, but as Malcolm Turnbull pointed out on SBS's Insight last night, a measly 2 Mbps, if 'synchronous' (equal upload and download speeds) is enough to permit standard definition video-conferencing.

He will use this kind of detail to argue that $2000 per capita to deliver 50 times that speed to 93 per cent of homes in the nation is a reckless and anti-competitive indulgence on the part of Labor. The current NBN rollout is supported by legislation that exempts it from ACCC scrutiny under the Trade Practices Act, and that prohibits Telstra using its HFC cable network to compete with the NBNCo.

That is in stark contrast with Labor's embrace of the competition principle across many of its other economic reforms – its rejection of calls for protection of markets such as agriculture, its on-market water buy-back plans, and its new pricing regime to assist public hospitals to compete for funds based on their actual service delivery.

The Coalition lost some of its pro-competition credibility last week with Joe Hockey's call to use "policy levers" to prevent banks setting their own rates, though this detour from core Liberal Party principles was ended in the party room yesterday, with Deputy Prime Minister Julie Bishop reportedly giving Hockey a firm reminder of where the Party stands on free markets.

But Malcolm Turnbull must take care not to swing too far away from Labor's fibre-based broadband plan. During the election campaign I argued that a broadband policy somewhere between the NBN Co and the Coalition's $6.7 billion plan might just have got Tony Abbott into the lodge.

Labor was proposing a Lamborghini, and Tony Smith and Andrew Robb fronted TV cameras to sell a second-hand Fiat. That not only lost the Coalition votes, but was pivotal and firming the support of Rob Oakeshott and Tony Windsor for Labor.

Turnbull must now get the sign-off from his party for a 'Malcolm in the middle' broadband plan. There are still juicy political points to be scored by pursuing a Productivity Commission cost-benefit analysis of the NBN, but Turnbull would be better served by letting that one die on the vine.

He should now focus his attention on a simple problem: fibre or copper, and in what proportions.

What has not been widely understood in the NBN debate is how broadband costs increase exponentially as the percentage of homes to be connected to fibre-optic cables is increased.

As a high-speed broadband roll-out in regional Britain shows (BT's lesson for Conroy's NBN, October 6), by connecting 50 per cent of homes to fibre, 40 per cent to copper and fibre (fibre to the curb) and 10 per cent to wireless broadband, the cost of the rollout comes in at between one-quarter and one-fifth of the per capita cost of the NBN.

And remember, that is a 'regional' roll-out connecting many isolated properties, whereas the NBN per capita cost is an average across a population that is 85 per cent urban.

There are compelling arguments for connecting every school, medical centre, business district (even a small cluster of businesses in a country town) and hospital to fibre, but what Turnbull will now have to argue is that many homes in Australia can achieve everything they want using a 'last mile' copper loop that may only deliver somewhere under 40 Mbps speeds, depending on how long, and in what condition, that last mile of copper is in.

Copper rotting in the ground is not an easy sell alongside 'future proof' fibre, but it has the huge cost advantage of already being in place. Add to that Turnbull's proposal for a private, but regulated Telstra wholesale company, the efficient use of existing HFC networks, a base-line 12 Mbps universal service obligation, and no curbs on competition whatsoever, and Turnbull really has something to sell.

So where does that leave Stephen Conroy?

Not as stranded as one may think. However, the threat of a Coalition policy that shames the NBN as a colossal, anti-competitive 'white elephant' should make him urgently revise the way he is selling his plan.

In fact, as time passes it's becoming clear that Stephen Conroy has done a poor job of selling the NBN. Even last night on SBS, as he and Turnbull wearily traded blows, he allowed the debate to remain centred around 'speed'.

That is the wrong approach. Yesterday I spoke at length to Doan Hoang, Professor of Computer Networks at UTS and director of the Advanced Research in Networking Group. While at pains to remain outside the political debate, Professor Hoang said it is almost nonsensical to discuss broadband speeds in isolation.

The 40 Mbps/10 Mbps copper/fibre offering in the UK sounded fine in principle, he said, but in practice speeds begin to drop based on the condition of the copper last mile, distance from the exchange and congestion on the network. Wireless, he said, is even worse – while it can promise 12 Mbps 'peak speeds' there is no guarantee that wireless networks will deliver that. Signal interference and congestion can slow wireless networking to a crawl.

Fibre-optic cabling, but contrast, is absolutely what Conroy makes it out to be – "future proof".

There was some scepticism when Conroy announced 1,000 Mbps would be achievable over the NBN Co network. In fact, says Hoang, the most anybody has fired down a single fibre is an astonishing 128,000 Mbps (128 Gbps). That's achievable only in a lab because of the cost of the 'multiplexers' sending and receiving the signal. However, whatever budget equipment NBN Co is installing now could, in theory, be upgraded in future to make fibre broadband effectively speed limit free.

But all of that is still missing the point says Hoang. Most business, health and even entertainment applications will use substantially less than the 100 Mbps that Conroy is trying to sell to the nation. But what really counts, it the total lack of congestion on the NBN Co fibre network. If you sign up to 25 Mbps, that is exactly what is achievable through your connection.

Hoang sees that as an exciting prospect that Australian businesses have not yet understood. He urges businesses to set up their own fibre-based experiments to begin testing business applications that they may want to use when fibre is more ubiquitous – only by doing it, he say, will they understand the way fibre-based networks will open new revenue streams.

So Stephen Conroy's shining vision is not dead. He just needs to start selling the real virtues of his network – total dependability and lack of congestion – rather than engaging in simplistic hand-to-hand combat with Malcolm Turnbull on 'speed'.

When Australians understand that, there may be more unanimous support for the idea of spending $2000 per head on the NBN.