Kiwis come clean on cabinet speeds

Statistics out of New Zealand and the United Kingdom show that vectoring may not the cure-all that the new Australian government hopes it will be.
Written by Phil Dobbie, Contributor on

One of the most telling facts in the Australian broadband landscape is that just over the water, the Kiwis already have fibre to the node (FttN), or '"to the cabinet", as they like to call it, and are now building fibre to the premises (FttP). In Australia, we're still arguing over what technology to use.

Britain sits in between, with OpenReach a long way down its FttN build, but with forecasts for FttP markedly downgraded. This different mix of technologies is creating varying expectations of speed, with Australia lagging sadly behind.

This week, TrueNet released its latest speed test data from over the ditch. It is an independent company contracted by the NZ Commerce Commission to provide monitoring of broadband performance. Volunteers perform the tests across a variety of networks, so that unlike most tests, the speed can be related to the technology, the provider, and what's promised in the plan.

When these figures are combined with data from the UK's regulator Ofcom, we quickly see how irrelevant ADSL2+ is becoming. Technology might always be able to squeeze more out of a copper line, but it seems that there's little more to be squeezed out of those longer runs. The only advances come from making the copper shorter.

It also shows that in its current form, VDSL is falling a long way behind a good fibre service. Snap (not shown in graph below) and Telecom in New Zealand average 20Mbps using VDSL. BT's "up to 38Mbps" plan averages between 31Mbps and 34Mbps.

Image: Phil Dobbie/ZDNet

Ahead of the pack is fibre, as one would expect. Snap is averaging 92Mbps on its fastest plan; Orcon is on 80Mbps. That's good, but FttN advocates might argue that it comes at a lot of extra build cost for such a small comparative advantage, particularly as VDSL speeds will be pushed up with vectoring. Already, BT's "up to" 76Mbps VDSL service is averaging 60Mbps.

In the midst of all of this sits hybrid fibre-coaxial (HFC). Vodafone in New Zealand rates badly, but it only recently acquired the network a few months ago. Whilst it rates worse than VDSL in the figures, TrueNet founder John Butt said the mean can be deceptive — half of its probes now report 100Mbps. In the UK, Virgin Media's "up to 100Mbps" service is hitting speeds close to the fastest fibre provider in NZ. Some are going even faster.

Yet, Australian cable seems to be a long way behind. Our latest ZDNet Speed Test data (from back in March) had cable averaging less than 20Mbps.

At the NBN Rebooted conference last week, Cisco's CTO Kevin Bloch was keen to make the point that HFC has a lot of life in it, with the DOCSIS 3.1 platform capable of speeds up to 10Gbps downstream and 1Gbps upstream. Of course, that'll be shared, so the theoretical maximum is pretty meaningless, but he made the valid point that upgrading the existing network will be quicker to deploy and faster to use than VDSL, as well as being a darned sight cheaper than building fibre to every premises.

Whatever solution is arrived at for the revamped NBN, it would make sense to include HFC, rather than closing the networks down — what industry consultant Dermot Cox calls "a destruction of capital". It's also clear that with ZDNet data showing Australian DSL speeds averaging 6Mbps, we have a lot of catching up to do, and making better use of the HFC networks could give us a much needed leg up.

Dermot Cox wrote to the Australian Competition and Consumer Commission in August (PDF), suggesting that HFC become a declared service, giving open access to retail competitors. He reckons 7 million Australians fall within the fibre footprint (it's hard to assess, because Telstra and Optus don't make the extent of their networks public knowledge). Those people, if they want faster speeds, really have no choice at present but to buy from one of these two companies — at least until the NBN is finished.

Meanwhile, as the Kiwis move to fibre, we are likely to edge slowly toward fibre to the node, whilst hoping that vectoring — a technology scarcely out of the laboratory — will push us closer to fibre-like speeds. The TrueNet figures show what a big job vectoring has if it is going to deliver speeds from the cabinet that get anywhere close to what's currently enjoyed with fibre or the UK experience of cable.

Incidentally, for those who think it would be great to have TrueNet-style reporting of our broadband speeds, it is looking for volunteer monitors. If you continually moan about the lack of reliability in speed reporting, now's your chance to make a difference.

Editorial standards