We are using the wrong broadband statistics

TrueNet is a New Zealand company that measures actual internet speeds from the premise to both local and international addresses. Pacific Fibre asked them to pull together from their data a chart showing the difference between local and international traffic speeds at peak times. (If there is an Australian company offering this service then please get in touch)

The left hand column reflects the speed from residences to a domestic NZ address. It’s this statistic (perhaps averaged out over the day, and measured from the server rather than the premise) that local and international surveys report when comparing Australia and New Zealand to the rest of the world. Akamai, Net Index, and Cisco for example, report on speeds from the premises to a local (inside NZ or Australia) server.

The right hand side column shows the effective speed from NZ premises to a server in Dallas, and thus includes the international link.

If we believe industry estimates over 80%+ of traffic to the home comes from offshore. Thus from the end user perspective the effective speed of the internet is generally much closer to the bar on the right hand side.

In the USA the left hand and right hand columns are one – as the vast majority of traffic is domestic. Similarly in Japan, Korea and, I take it, European countries like France – where the language dictates that most of the traffic is domestic. However when we compare average US speeds to speeds in Australia and New Zealand then the only fair way to do so is to use a measure of effective speed. That means we should be much lower in the international rankings, and thus should sound alarm bells.

Published by Lance Wiggs


4 replies on “We are using the wrong broadband statistics”

  1. Google Measurement Lab also does a great job here – and has powerful international comparisons.


    It works by the user running a browser based tests [a bit like, but more sophisticated than] speedtest.net – so gives comparable “real” results llike TrueNet


  2. Spot on. Combine that fact with the data caps and data costs most of us face with broadband in NZ and it’s a grim situation indeed.

    To elaborate, the slow international bandwidth and the cost of data mean that a whole range of network-based applications and services and aren’t feasible or aren’t cost-effective.

    My own company launched an automated cloud-based network backup service last year and didn’t bother to offer it to New Zealand customers because it was obvious that the data costs (and slow speed) would make access to this technology inaccessible to NZ businesses.


  3. This is why UFB will not infact turn out to be ultra fast.

    But most people will only realise this after the next election, when they finally get ‘UFB’ and use it themselves.


  4. Its also a case for more locally hosted material. My blog, and yours, and I’d bet many commercial sites are hosted offshore. So a fair chunk of that traffic may well be ,local information that has to take the long slow way home. What would be the effect of having better, more affordable, hosting in NZ?

    Do we have any stats on .nz domains that point to offshore servers and the scale of the traffic they generate?


Comments are closed.