Why copper and HFC can't save us

Of all the NBN confabulations and distortions, one of the worst is the ongoing suggestion that the copper network — if we just give it a few nips and tucks — is good enough to keep us basking in 12Mbps glory for many years to come.

Not to teach you to suck eggs, but remember that DSL technologies use specific frequency bands to transmit signals across standard copper wiring. The longer the distance the signal has to travel, the more that signal is attenuated and the slower the speed of the broadband; this is why, because I live nearly 5km as the crow flies from my local exchange, my ADSL2+ service was delivering just 2 to 3Mbps tops even though ADSL2+ is rated to 24Mbps. You will only get double-digit speeds if you live less than 3km from your local exchange; TPG, for one, offers a distance-versus-speed chart here, if you want to see how quickly speeds fall off.

Turnbull may have understood this, but he regularly misrepresents it when he says, as he did in a recent conference call with journalists, that fixing up broadband blackspots — which he identifies as the areas hobbled with Remote Integrated Multiplexers (RIMs) and pair-gain configurations and those areas housing "people living too far from exchanges to get satisfactory speeds" — can be accomplished with "quite a lot below $2 billion".

If you do a bit of playing around on Google Earth, it quickly becomes clear that there is something fundamentally wrong with the Coalition's maths and geography.

Sure, RIM and pair-gain configurations can be targeted and replaced with a government hand-out to Telstra, which should never have put them there in the first place. But I'd wager that the last group — people who live too far from exchanges to get satisfactory speeds — actually comprises quite a significant proportion of Australian households.

Thanks to ADSL2Exchanges.com.au, it's eminently possible to see just how much of a problem the copper network poses. Consider, for example, the city of Ipswich, Queensland, which I picked completely at random. Residents of Ipswich are served by one main exchange in the city's centre, the Brassall exchange 4.4km to the north, the Wallon exchange 9.6km to the west, the Amberley exchange 8.2km to the south-west, the Swanbank exchange 4.8km to the south and the exchange Bundamba, 5.8km to the east.

Telstra has, of course, sited its exchanges based on population considerations as well as the special high-density needs of locations like the RAAF Base Amberley, which lies just outside of Ipswich. Changing demographics have also changed early assumptions, with higher-density living and new population corridors putting previously unexpected pressure on exchanges. But if you do a bit of playing around on Google Earth, it quickly becomes clear that there is something fundamentally wrong with the Coalition's maths and geography.

Let's say, for rounding's sake, that everybody within 2km as the crow flies (remember that the length of copper between your house and exchange is longer than the as-the-crow-flies distance) can reach 12Mbps — meeting the Coalition's policy goal to provide "affordable broadband across the country". If we draw circles of approximately 2km radius around each exchange in the Ipswich area, we end up with a heck of a lot of areas where there is no colour. In other words, these areas are unlikely to receive 12Mbps in their wildest dreams, no matter what the Coalition does to boost their speeds. The mechanics of ADSL, and the copper-loop design choices Telstra made years ago, simply will not allow it.

Many residential areas in and around Ipswich are much more than 2km from local exchanges. (Screenshot by David Braue/ZDNet Australia)

Do the same in Turnbull's own electorate of Wentworth, and it's easy to see why he is so complacent: in densely-populated Sydney, exchanges are closer together and distance-generated attenuation would not seem to be such an issue, since you're rarely more than 2km from an exchange. Like Ipswich, Wentworth is serviced by six exchanges, at Vaucluse, Rose Bay, Bondi, Edgecliff, Waverley, Randwick. Plotting a 2km map around these exchanges reveals very few areas where the Coalition's dream could not plausibly come true.

It's hard to get more than 2km from an exchange in Malcolm Turnbull's electorate of Wentworth. (Screenshot by David Braue/ZDNet Australia)

So, perhaps, Turnbull's contention is based on the experience of his own electorate. Yet he should definitely expand his horizons: a similar map of coverage in the City of Onkaparinga, SA — one of the five NBN first-release sites — shows large areas outside our 2km boundaries. These areas may be able to get phone services at longer distances from the exchanges, but with peak theoretical speeds in the single-digit range there's a long way to go to fulfil the Coalition's dream.

The NBN will plug a lot of coverage holes throughout the City of Onkaparinga, SA. (Screenshot by David Braue/ZDNet Australia)

I tried the same trick in Darwin, which also has six exchanges covering its closest population centres. Again, we have massive areas where 12Mbps is simply not going to be available. The further out you go from any of our CBDs, the smaller the proportion of our country that's covered with colours. The failing here is not the copper network, but the realities of ADSL technology which might as well stand for "Additional Distance Slows Line".

Darwin has grown since this imagery was taken in 2006, but Telstra's exchange network only covers parts of the city well enough to deliver 12Mbps ADSL2+ services. (Screenshot by David Braue/ZDNet Australia)

These distance realities mean that no matter how many RIMs or pair-gain systems we replace, Coalition strategies are going to do basically nothing to improve service for people that already have ADSL but are simply finding that it's under-delivering. Unless Malcolm Turnbull is going to double or triple the density of Telstra's current 3000-odd exchanges around the country that is not going to change.

These realities also bode poorly for new DSL-based technologies that are now claiming to push hundreds of megabits through existing lines. Nokia Siemens Networks, for example, recently debuted noise-cancelling technology that it says can deliver 825Mbps over — wait for it — 500m of copper; this came hot on the heels of an announcement last year that NSN could deliver 25Mbps at a distance of 1.5km. Alcatel-Lucent announced a similar technology earlier in the year to provide 300Mbps over 400m of multiple aggregated copper connections.

Note first of all that these are theoretical maximums, second of all that they fall well short of resolving coverage issues past 2km, and third of all that they seem to require the use of multiple copper lines. Does your house have multiple copper lines running into it? Mine certainly doesn't.

In other words, DSL works great in laboratories over short runs of copper, but our copper network is not built using short runs of copper.

The hybrid-fibre coaxial (HFC) myth

Turnbull's other suggestion is that "we" should capitalise on Telstra's and Optus' existing cable networks to deliver 100Mbps speeds wherever possible. "You can deliver 100Mbps over the HFC network now, if it's tuned up," he said in his recent press conference. "The logical thing that would happen under the proposal I've got now, is that both of those networks would be switched up to 100Mbps."

Coalition strategies are going to do basically nothing to improve service for people that already have ADSL but are simply finding that it's under-delivering.

The thing is: the HFC networks only cover a small and carefully defined part of Australia's geography. Their networks are closed to competitors, under-subscribed and hardly at risk of being upgraded after they haven't been voluntarily extended in the past 10 years. Turnbull seems to be suggesting that the government bankroll the cost of Telstra and Optus to finish upgrading their HFC networks — yet isn't the Coalition's whole telecoms platform based on the idea that private investment will ultimately provide the services the market wants? If this is the case, why would he be stumping for subsidies to a pair of capital-city networks that are probably the best example of perfect competition the country has? Wouldn't that be a concession that the private sector just isn't interested — even in investing in wholly-owned networks over which they already have control?

Turnbull's conventional arguments against the NBN are based on a series of ever more tenuous assumptions. So when he says "if you could deliver, for example, nationwide 12Mbps at a relatively modest cost compared to the NBN, what is the additional utility benefit value of going from 12Mbps to 100Mbps?" he's deflecting the argument to be one of positioning one speed against another, when this is a false dichotomy.

That dichotomy is designed to distract from the reality: that you cannot actually deliver nationwide 12Mbps at a relatively modest cost compared to the NBN. Filling in the gaps I've highlighted would require not only the Coalition's planned RIM-replacement work, but extensive supplementation of the existing exchange infrastructure with new exchanges; billions of dollars to reroute homes' copper services from their existing distant facilities to the new sites; and all this to bolster a copper infrastructure that not even Telstra wants anymore.