Broadband performance measurement specialist TrueNet has branched into evaluating mobile broadband performance and claims to have found some surprising disparities.
The provider that clocked up the best performance in web-browsing in a series of tests around Wellington and the Hutt Valley was the worst performer when it came to bulk downloading of data.
The best download performer of the three mobile providers came second in the browsing-speed stakes, but was a long way behind the leader (see graph below).
On the basis of these limited tests, before choosing a provider users should consider what is likely to be their predominant use of mobile broadband – bulk downloading or browsing – and also, perhaps, the reception characteristics of your most frequent location, says TrueNet director John Butt.
TrueNet’s preliminary report does not identify the carriers individually making it, for the present, of limited value to those wanting to choose a provider. The company will reveal the individual carriers and test results following more comprehensive trials, “in the next few weeks”, Butt says.
Results were variable, TrueNet’s report states. Top speed for all three carriers was around 7 Mbps, but clearly the worst performers turned in more individual slow results. “All three carriers manage to have a similar proportion of tests where their speeds are more than say 3Mbps, but some have a lot more very slow tests,” the report states.
TrueNet is a joint venture of software developer Catalyst with a 65 percent shareholding and Jonette Consulting, run by John Butt, an expert in broadband performance.
Jonette Consulting previously measured stationary broadband performance for the Commerce Commission’s quarterly broadband quality reports up until mid-2010. The company stopped this work when planning the TrueNet venture, because of a potential conflict of interest, Butt says.
Measurements are averaged from a number of monitoring points, to date mostly in central Wellington and Lower Hutt.
A second series of measurements, now under way, will compare performance of the three carriers at the same single point and time, and present them as a table with a column for each monitoring point and a row for each carrier.
Trying to take all the other variables out of the equation to arrive at a valid comparison is difficult, Butt says.
One technique used in the published report is to divide the transmission speed by a measure of the signal strength. This avoids giving a misleading low rating when the low speed is owing to a weak signal, which may be a transitory problem or an effect of a particular site’s geography, he says.