I've already tested how 3G broadband works between on trains Melbourne and Sydney and Sydney and Brisbane. So it was inevitable that I'd have to try a similar trick travelling between Melbourne and Adelaide. Here's what I learnt.
First things first: while the principles are the same, these are not a set of directly comparable tests. The Melbourne-Sydney route was designed to see how Telstra's standard Next G service worked. (The answer? Poorer coverage than you'd expect.) The Sydney-Brisbane route tested, in stages, how Telstra's DC-HSPA+ enhanced Ultimate service worked. (The answer? Often slower than you'd expect for a premium service.)
This time around, I was testing Telstra's standard Next G service via a USB modem on my PC, and comparing the availability of Optus 3G service on my BlackBerry at the same time. (I abandoned plans to test Vodafone as well after my SIM ran out of credit early in the journey; while Vodafone has said it is improving its network, I suspect it would have been the loser in this test given its relatively poor regional coverage.)
I wasn't trying to test broadband continuously throughout the journey, as I've often done before. In part, this was because both previous tests suggested that this fails with astonishing frequency. I'd also planned to devote the day to NaNoWriMo editing, and I didn't want consistent interruptions to check Internet connectivity. So I decided to run speed tests for Telstra each time we stopped at a station, which would mean by definition we were in a relatively populated area.
The most surprising lesson this time around, as I foreshadowed with a brief post on Friday, was that in a couple of rural areas, Next G actually proved to be completely unavailable while Optus was offering a functional signal. I think this is worth pointing out because Telstra has very successfully managed to create a perception that it will always win when it comes to regional coverage, and I don't think that perception always matches with reality when you actually put it to the test.
That doesn't mean Next G has lousy rural coverage; it means that you always need to check before signing up if access is important to you in a given area. And note that this isn't an issue about speed as such: it's an issue about whether you get a signal of any description whatsoever.
That incident aside, the speed results (via Speedtest.net) in rural locations varied quite widely, as you can see below (click for a larger version):
We didn't actually stop at Nhill or Bordertown stations (that only happens when passengers are booked to alight or board), so those tests do represent on the move data where a signal was maintained. That wasn't always going to happen: we'd barely left Dimboola station when the network disappeared completely. The Bordertown results are unusual in the closeness of upload and download speeds, but showed the same pattern when I tried a repeat test.
While the ping times are on the low side, on the whole the speeds didn't vary hugely from what Telstra offers in city areas. They're certainly more than enough to allow anyone trying to do essential tasks to get through work, especially if you're using a smart phone that can deal with intermittent connectivity more intelligently than a PC.
No 3G broadband network is perfect, and that is the most important lesson I've learnt this year while roaming the country testing their performance. I started writing the draft of this article on the platform at Sydney's Central Station -- the absolute centre of the nation's biggest city -- and it took 10 minutes to persuade an Optus 3G modem to even work. I wouldn't want to ever lose the convenience of having wireless broadband, but anyone who thinks it's a comparable substitute for wired Internet access in terms of reliability is delusional.
Lifehacker Australia editor Angus Kidman is experiencing mixed feelings about Adelaide public transport. His Road Worrier column, looking at technology and organising tips for travellers, appears each week on Lifehacker.