How long’s a piece of copper?

The answer to that question is important. The bandwidth that the Coalition will be able to deliver to customers over its proposed fibre to the node network is highly dependent on the length of copper from the node to the customer, but neither it nor anybody outside Telstra seems to know what the average distance from nodes in the current network is.

In the FAQs on the policy published by shadow communications minister Malcolm Turnbull on his blog, in response to the question: “How long will the copper loop be?” he says: “The last serious plans for FTTN in Australia in 2007 and 2008 had maximum loop lengths of between 750m and 800m. That would mean that minimum speeds of 25Mbps are more than feasible. Recent trials by Alcatel have shown that even at lengths of 750m they are getting speeds in excess of 50Mbps so the technology is improving all the time.”

Note that he is not answering the question, because he does not have the answer. When I asked his office about this I received the response: “We can’t design the network from Opposition – all we can say is what very detailed plans in the past have relied on and set speed targets that network architecture specifications need to be built around (and not the other way around).”

That’s all very well but shortening the length of copper from the present locations of the street corner pillars at which either passive splitters for the FTTH network or opto-electronic converters in an FTTN network could be located with relative ease would be a costly exercise.

According to Paul Fletcher, in his book ‘Wired Brown Land’, this 800m figure comes from plans submitted by Telstra to the Coalition Government in 2007, just before the ALP came to power and embarked on its FTTN plan (aka NBN Mark 1). Fletcher said that Telstra was talking about average loop lengths of 800 metres and he said that this figure represented a significant reduction from the 1.5kms in its first FFTN proposal in 2005.

Fletcher pointed out, rightly, that the cost to reduce the loop length would be significant and he suggested such a projected cost would have to be disclosed to the Stock Exchange. He had another explanation: “The most likely explanation is that Telstra did not really change its plans at all; it merely changed what it disclosed publicly.”

My next port of call in search of this elusive figure was Brian Beckor, founder and CEO of Callpoint Spatial, a company that has provided advice to numerous telcos on the topology of the telephone network. He was not very helpful.

“Can anybody get their hands on information on the copper runs and use it publicly? I doubt it,” he said. He explained that bidders for the ALP’s FTTN network had been provided with sample data for 143 of the more than 5000 exchange areas in the Telstra network, but under strict non-disclosure agreements.

Not that Beckor hasn’t tried to find an answer, and he said an estimation of sorts was possible. “We don’t have the co-ordinates of Telstra’s pillars. We tried to get a hold of this about four years ago and came up with all sorts of ideas including Google satellite images, crowdsourcing - people taking photos - but we could not do enough of it in a short enough time to deliver the goods.”

However he could come up with an estimate. He has information on the boundaries of all distribution areas every one of which must have at least one pillar. If he were to assume one pillar at the centre of each distribution area, it would be possible to calculate a figure for the mean and median distance to premises in that area.

He also offered one figure that calls into question the Coalition’s claims for the bandwidth it will be able to deliver over its planned FTTN network. In his FAQs Turnbull says: “Recent plans for FTTN networks in Australia indicated that around 50,000 nodes would be needed (many of them in the basement of apartment blocks).”

But, according to Beckor there are 72,322 distribution areas in the country, which means at least 73,322 pillars. While that number does include those in rural and regional areas that will not be served by FTTN or FTTH, it does not include any in the basements of apartment blocks.

Fewer nodes than the number of pillars in the current network means longer copper lengths, and the bandwidth that can be delivered over copper decreases rapidly at distances in excess of 1km. Over to you Mr Turnbull.

This article first appeared on iTWire, Australia's leading independent IT&T news and information source.