August Broadband report, including a multiple website comparison

Forums › New Zealand Broadband › August Broadband report, including a multiple website comparison

JohnButt

374 posts

Ultimate Geek

Trusted

#109478 20-Sep-2012 13:55

August report with ISPs in Two Distinct Groups

The TrueNet Time-of-Day Speed measure reports that, during peak load periods, five ISPs met the TrueNet standard of 95% of their fastest speed. This is an improvement on 8 months ago when only Snap and Telecom met the TrueNet standard.

For the first time TrueNet is publishing a Surfing the Internet by Time-of-Day analysis, which has Snap & Telecom alone achieving better than 90% of maximum performance.

www.truenet.co.nz

Ragnor

8085 posts

Uber Geek

Trusted

#688945 20-Sep-2012 17:02

John you tease at least post a click-able link to the article
http://www.truenet.co.nz/articles/august-2012-broadband-report

Sharesies

Trade NZ and US shares and funds with Sharesies (affiliate link).

JohnButt

374 posts

Ultimate Geek

Trusted

#688997 20-Sep-2012 18:09

Apologies Ragnor, I was posting while mobile.

Thanks for the correction, but since I stuffed up, here is some more detail just for Geekzoners -

Actually I am looking for feedback on understandability, the concept of independent spots is difficult to get across, a 3D view looks better but the data is harder to read.

The chart should be self explanatory to be a good chart. Happy to answer questions to explain further, but lets see how well it is understood from just the chart.

Ragnor

8085 posts

Uber Geek

Trusted

#688999 20-Sep-2012 18:17

First impression: The dotted lines between two points for the same bank across ISP's makes no sense, there is no actual link. If you remove those it will make the graph clearer imo

NonprayingMantis

6434 posts

Uber Geek

#689003 20-Sep-2012 18:23

JohnButt: Apologies Ragnor, I was posting while mobile.

Thanks for the correction, but since I stuffed up, here is some more detail just for Geekzoners -

Actually I am looking for feedback on understandability, the concept of independent spots is difficult to get across, a 3D view looks better but the data is harder to read.

The chart should be self explanatory to be a good chart. Happy to answer questions to explain further, but lets see how well it is understood from just the chart.

what is interesitng to me is the juxtaposition of these results (above) with the ones on the website

the chart above suggests that Telecom on average has the poorest performance of that group when looking at bank websites, but VF, TCL, and Slingshot all have very good performance

However when you look at this chart

it suggests that, overall, Telecom is roughly at parity with VF and TCL, but slingshot, one of the top performers for 'bank' access, is way worse, posting times around triple that of TC,VF and TCL.
Whilst this presumably includes sites other than banks, it is still is a complete reversal of fortune for both Telecom andslingshot compared to the first chart, which seems odd

any idea why that might be?

insane

3170 posts

Uber Geek

ID Verified

Trusted

#689011 20-Sep-2012 18:28

Will be down to their peering, or rather lack thereof.

Ragnor

8085 posts

Uber Geek

Trusted

#689023 20-Sep-2012 18:44

insane: Will be down to their peering, or rather lack thereof.

It will be interesting to see what Vodafone does once they consume Telstraclear. I presume bean counters will want to preserve transit revenues so Vodafone might regress to not openly peering at APE/WIX etc (like Telstraclear and Telecom).

JohnButt

374 posts

Ultimate Geek

Trusted

#689025 20-Sep-2012 18:45

Do you think that is due to the difference of routes? I notice some issues that look like "bufferbloat" on some, but not all Bank downloads. It is particularly evident on the Kiwibank downloads.

The dotted lines is the way I thought you could show the relationship between websites, when I showed only dots, it became difficult to see - as it was with 3D cones

Just a point I missed on the first post, these tests could only be done on the 5GB/month capped volunteers lines, which do make up over 50% of our probes. However, this is why the smaller ISPs dropped out
A hint about the impact of caps? eg we have nearly all Slingshot connections on 5GB/month, but very few Xnet connections (4), despite the number of total probes being similar.