Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.


JohnButt

374 posts

Ultimate Geek

Trusted

#220128 27-Jul-2017 12:50
Send private message

MyRepublic appear to be going through a lot of change in performance.  Better backhaul or CDN access?

 

TrueNet monthly report for June 2017 focusses on the change in webpage average speeds where we download over almost 50 pages on all technologies.

 

See it here

 

https://truenet.nz/story/2017/07/june-2017-urban-broadband-report 


Create new topic
sbiddle
29208 posts

Uber Geek

Moderator
Trusted
Biddle Corp
Lifetime subscriber

  #1832215 27-Jul-2017 13:22
Send private message

What are you doing about all your (obviously flawed) data that's showing multiple providers with peak time speeds well in excess of 100Mbps?

 

https://truenet.nz/story/2017/07/june-2017-urban-broadband-report#Speed


JohnButt

374 posts

Ultimate Geek

Trusted

  #1832326 27-Jul-2017 15:19
Send private message

sbiddle:

 

What are you doing about all your (obviously flawed) data that's showing multiple providers with peak time speeds well in excess of 100Mbps?

 

https://truenet.nz/story/2017/07/june-2017-urban-broadband-report#Speed

 

 

I have spent the last day checking these results - our analyst had already checked them 2 weeks ago, so this is a waste of time, but I have found out about an interesting feature of the NZ fibre network.

 

I can confirm our Fibre 100Mb/s chart is accurate.  There is NO FLAW.


 
 
 
 


chevrolux
4592 posts

Uber Geek

Trusted

  #1832346 27-Jul-2017 15:53
Send private message

I'm trying to understand this too. How can a panelist who has said they buy a 100Mbps service, and then receive in excess of 200Mbps, be used as proper data in these graphs?

 

Also, I assume it's up to the end user to tell you what their plan is. So they could quite easily just forget to update their profile after a plan change? And then there is the Vodafone data.. isn't the base cable plan these days 130/10 - which would line up with their results of around 120. Just doesn't seem right to have it on a graph "Median of Peak Speed for 100Mb/s Fibre services". Again, its up to the user to tell you their plan right? I bet no one gets confused with the "FibreX" plans.... 

 

I would expect to see results hovering around the 100-110 to account for the "over-provisioning" (like Chorus' right performing). But if they are consistently hitting speeds well above what they buy, how can that count as proper data - there is clearly just a provisioning issue. And I would say wrong speed allocations happen irregularly and are not "quite common".


Talkiet
4470 posts

Uber Geek

Trusted

  #1832347 27-Jul-2017 16:01
Send private message

I'm pretty sure I've worked out what's happened with the results. IF I am correct then they are accurately collected results for this narrow test scenario. Whether they could be interpreted as representative of real world experiences in general is another matter entirely. 

 

John knows my number. Happy to discuss if he'd like to call.

 

Cheers - N





--

 

Please note all comments are the product of my own brain and don't necessarily represent the position or opinions of my employer, previous employers, colleagues, friends or pets.


michaelmurfy
/dev/null
9546 posts

Uber Geek

Moderator
Trusted
Lifetime subscriber

  #1832349 27-Jul-2017 16:02
Send private message

Looks a bit flawed to me also. I can also see MyRepublic is excluded from quite a few of the charts. How do we know that it isn't simply the test probes having their connection moved to 200/20?





Oblivian
4232 posts

Uber Geek


  #1832357 27-Jul-2017 16:07
Send private message

michaelmurfy:

 

Looks a bit flawed to me also. I can also see MyRepublic is excluded from quite a few of the charts. How do we know that it isn't simply the test probes having their connection moved to 200/20?

 

 

Speedtest 2.0?


ubergeeknz
3344 posts

Uber Geek

Trusted
Vocus

  #1832361 27-Jul-2017 16:11
Send private message

A lot of the charts also just have a line for "fibre" ... when some are 30Mbps plans, some gigabit (or whatever we're allowed to call it now), etc, this doesn't seem to provide a meaningful comparison between the same service on two providers, what does it tell us?


 
 
 
 


JohnButt

374 posts

Ultimate Geek

Trusted

  #1832381 27-Jul-2017 16:41
Send private message

To help here:

 

 

 

1. Try not to shoot the messenger - we just test and report.   I can confirm that not only are the panelists confirming that they have bought 100, but also I have checked with what was happening in a number of networks in June.  

 

2. The test is not a "Narrow test example" it is a simple download test and we do a number of test of varying sizes to provide confirmation, including 3 separate tests from speedtest.net servers.

 

 

 

BTW, on Average Website Speed charts we mix all Fibre >=100Mb/s together, because if we showed 100, 200 and GigE together in one chart you will struggle to see the difference.


Talkiet
4470 posts

Uber Geek

Trusted

  #1832391 27-Jul-2017 16:52
Send private message

JohnButt:

 

To help here:

 

 

 

1. Try not to shoot the messenger - we just test and report.   I can confirm that not only are the panelists confirming that they have bought 100, but also I have checked with what was happening in a number of networks in June.  

 

2. The test is not a "Narrow test example" it is a simple download test and we do a number of test of varying sizes to provide confirmation, including 3 separate tests from speedtest.net servers.

 

 

 

BTW, on Average Website Speed charts we mix all Fibre >=100Mb/s together, because if we showed 100, 200 and GigE together in one chart you will struggle to see the difference.

 

 

So you believe that users would be able to sustain those reported speeds over several seconds, or even minutes?

 

Cheers - N

 

 





--

 

Please note all comments are the product of my own brain and don't necessarily represent the position or opinions of my employer, previous employers, colleagues, friends or pets.


Jase2985
9935 posts

Uber Geek

Lifetime subscriber

  #1832394 27-Jul-2017 16:55
Send private message

im just curious when you are going to do larger file sizes, like 50mb say. 5mb is nothing these days on the likes of a 200mbps connection, it doesn't even have time to ramp up fully.


JohnButt

374 posts

Ultimate Geek

Trusted

  #1832752 28-Jul-2017 11:35
Send private message

Jase2985:

 

im just curious when you are going to do larger file sizes, like 50mb say. 5mb is nothing these days on the likes of a 200mbps connection, it doesn't even have time to ramp up fully.

 

 

We do larger files to test rampup, 5MB and 14MB.  Rampup is typically good for just a 2MB file for 100Mb/s with only infrequent errors caused by sampling, 5MB is rarely short of two quartiles at the same speed.  We have also run a few tests on 100MB files and found them no different in overall results from our 5MB files.  I believe we should also not load the NZ network with enormous data queries just for testing, BTW we do each test 720 times per probe per month - that can be a lot of data.

 

The bigger question, and one which I am interested in a debate over, is "What is the real speed?  Does it include the impact of rampup or is it just bragging rights to the biggest number?"  What is the point of knowing the biggest speed?

 

Why I put it that way is that some ISPs have significantly slower rampup than others, and that impacts on performance, and it may be especially so for browsing.  

 

Maybe we should report rampup separately from speed?  Maybe we should wait for the burst that shows the biggest possible number?  I would appreciate hearing opinions


JohnButt

374 posts

Ultimate Geek

Trusted

  #1832755 28-Jul-2017 11:36
Send private message

Talkiet:

 

JohnButt:

 

To help here:

 

 

 

1. Try not to shoot the messenger - we just test and report.   I can confirm that not only are the panelists confirming that they have bought 100, but also I have checked with what was happening in a number of networks in June.  

 

2. The test is not a "Narrow test example" it is a simple download test and we do a number of test of varying sizes to provide confirmation, including 3 separate tests from speedtest.net servers.

 

 

 

BTW, on Average Website Speed charts we mix all Fibre >=100Mb/s together, because if we showed 100, 200 and GigE together in one chart you will struggle to see the difference.

 

 

So you believe that users would be able to sustain those reported speeds over several seconds, or even minutes?

 

Cheers - N

 

 

 

 

 

 

Yes, although only while they have that service.


Create new topic




News »

HP unveils new innovations for businesses adapting to rapidly evolving workstyles and workforces
Posted 17-Sep-2020 15:36


GoPro launches new HERO9 Black camera
Posted 17-Sep-2020 09:45


Telecommunications industry launches new 5G Facts website
Posted 17-Sep-2020 07:56


New Zealand ranks 3rd in world in GSMA index
Posted 15-Sep-2020 10:13


Trend Micro Security Suite adds web monitoring to prevent identity theft
Posted 14-Sep-2020 15:37


NVIDIA to acquire Arm for US$ 40 billion
Posted 14-Sep-2020 12:27


Epson launches its next gen A3+ colour EcoTank multi-function printer
Posted 10-Sep-2020 16:08


Sony launches three new native 4K SXRD home cinema projectors
Posted 9-Sep-2020 18:00


Catalyst Cloud brings Kubernetes-based open-source web hosting solution to market
Posted 9-Sep-2020 17:54


Verizon Connect eyes further growth in New Zealand
Posted 8-Sep-2020 09:26


PNY launches XLR8 gaming NVIDIA GeForce RTX 30 series powered by the all-new NVIDIA Ampere architecture
Posted 3-Sep-2020 16:39


NVIDIA delivers greatest-ever generational leap with GeForce RTX 30 Series GPUs
Posted 3-Sep-2020 16:17


Weta Digital advances visual effects and animation in the cloud with AWS
Posted 2-Sep-2020 17:09


Kiwrious lab-in-the-pocket kit designed for schoolchildren
Posted 28-Aug-2020 09:03


Fitbit introduces Sense, its most advanced health smartwatch
Posted 26-Aug-2020 10:14



Geekzone Live »

Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.


Support Geekzone »

Our community of supporters help make Geekzone possible. Click the button below to join them.

Support Geezone on PressPatron



Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.

Alternatively, you can receive a daily email with Geekzone updates.