Speed test: Is Turnbull's solution quick enough?

It will be no surprise when Malcolm Turnbull announces a plan reliant on fibre to the node (FttN) and continued use of the Telstra and Optus hybrid fibre-coaxial (HFC) networks this week. The logic is sound: Make use of what infrastructure we have, and avoid the slow, expensive process of having to hook up each individual premises. The downside is, if you want the fastest speeds, you'll choose the HFC option, which could still mean hooking up your home using coaxial cable rather than a pure fibre solution. The hard work will still be done, only with a technically inferior option.

All options have their own theoretical top speed. Telstra promotes its cable services at speeds up to 100Mbps, and it could go higher — DOCSIS3.1 can, supposedly, offer a capacity of 10Gbps downstream. The fact is, though, that few are getting anywhere near those speeds.

Data from the ZDNet Broadband Speed Test (of 11,000 records of HFC users since September last year) shows that only 8 percent achieved speeds of over 40Mbps, and 30 percent didn't even reach 10Mbps. The vast majority is achieving speeds of less than 20Mbps, even allowing for Telstra's DOCSIS upgrades. Imagine the impact of putting even more traffic on that network.

Admittedly, there is a similar pattern with fibre users. Of the 1,400 tests over the same period, only 12 percent achieved speeds of over 40Mbps, but consumer pricing will have had a bigger part to play here. Most low-end fibre plans place a top limit at 12Mbps, to reflect NBN Co's wholesale pricing, so you could argue that this will artificially push more users down to lower-speed plans. Despite that, fibre users are generally getting a better experience — as you'd expect — and the network is built to withstand many more users, ultimately choosing faster plans.

DSL speeds remain well below HFC and fibre. No surprises there. ZDNet's Broadband Speed Test results for DSL have consistently averaged around the 6Mbps mark over the last eight months. 4G is providing speeds averaging one third faster. If you were to look at it simplistically, and ignore the relative download levels of fixed versus mobile users, you might conclude that the easiest way to improve speeds is to move DSL users onto wireless. As the data demonstrates, though, 4G speeds are already starting to slow as more people pile onto the network.

The Coalition's plan is to improve DSL speeds with fibre to the node — fibre to a cabinet in the street, and VDSL from there to your home. VDSL2 speeds can achieve 100Mbps, but only if you look out of your window and see a cabinet hanging off a telegraph pole. For anyone more than 500 metres from a cabinet, the speed quickly drops below 38Mbps. And, of course, it's heavily dependent on the quality of the copper coming into your home and your distance from the cabinet.

The Coalition could argue that its solution would still place people one or two brackets from the left side of our first graph — precisely where most people are with HFC and cable. It's a fair point, but one that ignores the future potential. Our graph shows where people are now; do we really want an infrastructure that can only cope with present-day demand?