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.


View this topic in a long page with up to 500 replies per page Create new topic
1 | ... | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19
SiliconAudio
96 posts

Master Geek


  #2104031 9-Oct-2018 17:14
Send private message

michaelmurfy:

 

Just something simple too - I changed a bit on my network and noticed my probe had stopped testing and had to reboot it. It may pay to do a quick "turn it off and on again" to see if this fixes it. Your speeds are looking incredibly low compared to the other 2degrees nodes so makes me think there is either something amiss on your network or you've got a faulty probe or a probe on the wrong testing profile.

 

 

I rebooted everything last night, including my ONT and router and even did a hardware reset on the SamKnows box.  Nobody has been home today and here's all it's captured:

 

Click to see full size

 

Not sure what's up, so I have reached out to SamKnows community support.  I have an Ubiquiti Edgerouter Lite with hardware offloading enabled for port forwarding, VLAN, NAT, etc. but otherwise a very simple setup.  Again, speedtests from this router show up to 800 Mbps down locally.  I have a few smart devices like a Google home, Philips Hue lighting, Smanos security system, but they generate absolutely minimal traffic during the day.

 

I'll report back if/when SamKnows tech support come back to me.  At this point, my unit is contributing no useful data, so if they can't figure this out, I will just pull it and get on with my life.


michaelmurfy
/dev/null
9641 posts

Uber Geek

Moderator
Trusted
Lifetime subscriber

  #2104050 9-Oct-2018 17:59
Send private message

@SiliconAudio My box has done 8 tests so far today so I'd say yours is on the wrong ISP profile. I'm sure they'll get this updated. Also, my network is in no means quiet (heaps of IoT, Netflix due to sick flatmate etc).





 
 
 
 


SiliconAudio
96 posts

Master Geek


  #2104053 9-Oct-2018 18:11
Send private message

michaelmurfy:

 

@SiliconAudio My box has done 8 tests so far today so I'd say yours is on the wrong ISP profile. I'm sure they'll get this updated. Also, my network is in no means quiet (heaps of IoT, Netflix due to sick flatmate etc).

 

 

It certainly lists the correct ISP and product in the portal, but I'll wait and see what they come back with.  I looked at the stats on my router and it's only been doing a few Kbps with the odd intermittent momentary peak at maybe 5 Mbps - certainly nothing significant, so I know there's no unexplained/unexpected traffic causing this.


sidefx
3510 posts

Uber Geek

Trusted

  #2105672 10-Oct-2018 15:44
Send private message

Mine arrived about 2 weeks ago, but only had a chance to pick it up from nz post centre a couple of days ago.  Will post some graphs when it's had a bit of time to collect a bit more data.  Is there a list of the MLab locations it tests against?  Some are obvious like Tokyo... others not so much. 





"I was born not knowing and have had only a little time to change that here and there."         | Electric Kiwi | Sharesies
              - Richard Feynman


SiliconAudio
96 posts

Master Geek


  #2106323 11-Oct-2018 15:31
Send private message

OK, so I seem to be getting more regular results now, but the speeds are all over the place, and patently incorrect.

 

Here's what speedtest.net gives me:

 

 

Here's what I'm getting from the SamKnows portal:

 

 

  • How can the international speed be faster than the domestic speed as it is between 16:00 and 20:00?
  • SamKnows domestic speed never above 280 Mbps, whereas speedtest to a national server is almost 750 Mbps, easily repeatable.
  • The speedtest.net result is from a laptop actually plugged into the SamKnows box, so this is not a problem on my LAN or gateway router.

I can't help but think the SamKnows box is a piece of junk.  If other people are collecting junk data with these boxes, the validity of the whole study comes into question.


hio77
'That VDSL Cat'
12626 posts

Uber Geek

Trusted
Subscriber

  #2106334 11-Oct-2018 15:50
Send private message

SiliconAudio:

 

  • How can the international speed be faster than the domestic speed as it is between 16:00 and 20:00?
  • SamKnows domestic speed never above 280 Mbps, whereas speedtest to a national server is almost 750 Mbps, easily repeatable.
  • The speedtest.net result is from a laptop actually plugged into the SamKnows box, so this is not a problem on my LAN or gateway router.

I can't help but think the SamKnows box is a piece of junk.  If other people are collecting junk data with these boxes, the validity of the whole study comes into question.

 

 

I wouldn't really go that far, but what i would say is SamKnows has come out without really too much consideration to the networks it's testing to.

 

 

 

EG. REANNZ, they posted on nznog that they were unaware the traffic was going to be as heavy, and thus were hitting Far heavier than expected traffic in peak times..





#include <std_disclaimer>

 

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

 


dfnt
1203 posts

Uber Geek

Lifetime subscriber

  #2106345 11-Oct-2018 15:57
Send private message

hio77:

 

SiliconAudio:

 

  • How can the international speed be faster than the domestic speed as it is between 16:00 and 20:00?
  • SamKnows domestic speed never above 280 Mbps, whereas speedtest to a national server is almost 750 Mbps, easily repeatable.
  • The speedtest.net result is from a laptop actually plugged into the SamKnows box, so this is not a problem on my LAN or gateway router.

I can't help but think the SamKnows box is a piece of junk.  If other people are collecting junk data with these boxes, the validity of the whole study comes into question.

 

 

I wouldn't really go that far, but what i would say is SamKnows has come out without really too much consideration to the networks it's testing to.

 

 

 

EG. REANNZ, they posted on nznog that they were unaware the traffic was going to be as heavy, and thus were hitting Far heavier than expected traffic in peak times..

 

 

What'd REANNZ expect with all these gig connections running speedtests all the time..


 
 
 
 


hio77
'That VDSL Cat'
12626 posts

Uber Geek

Trusted
Subscriber

  #2106376 11-Oct-2018 16:02
Send private message

dfnt:

 

hio77:

 

SiliconAudio:

 

  • How can the international speed be faster than the domestic speed as it is between 16:00 and 20:00?
  • SamKnows domestic speed never above 280 Mbps, whereas speedtest to a national server is almost 750 Mbps, easily repeatable.
  • The speedtest.net result is from a laptop actually plugged into the SamKnows box, so this is not a problem on my LAN or gateway router.

I can't help but think the SamKnows box is a piece of junk.  If other people are collecting junk data with these boxes, the validity of the whole study comes into question.

 

 

I wouldn't really go that far, but what i would say is SamKnows has come out without really too much consideration to the networks it's testing to.

 

 

 

EG. REANNZ, they posted on nznog that they were unaware the traffic was going to be as heavy, and thus were hitting Far heavier than expected traffic in peak times..

 

 

What'd REANNZ expect with all these gig connections running speedtests all the time..

 

 

https://list.waikato.ac.nz/pipermail/nznog/2018-September/022888.html

 

 

 

 

I thought I'd share what we're seeing hosting the M-Lab [1] nodes in NZ
that the new Comcom/Samknows [2] broadband testing are using for some of
their tests.

 

For a start, it looks like the testing has been ramping up over the last
week (see attached graph), in particular during the evening (6pm-midnight).
I had assumed the testing would be evenly spread out throughout the day,
but it looks like they're concentrating the testing during the busy evening
period.

 

 

 

 

 

It's worth noting we've been hosting M-Lab nodes for around 5 years now, but no one from Comcom or Samknows has spoken to us about the testing so it's all come as a bit of a surprise to us. If you have any interesting information to share, we'd like to know more :)

 

 

 

 

If you google "speedtest" the built in google speedtest does hit the mlab servers.





#include <std_disclaimer>

 

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

 


Djmixerdomo
241 posts

Master Geek

Subscriber

  #2106382 11-Oct-2018 16:13
Send private message

hio77:

 

If you google "speedtest" the built in google speedtest does hit the mlab servers.

 

 

That speed testing service is all over the show.

 

I just did five tests and got some interesting results:

 

Down / Up Mbps

 

77 / 92

 

104 / 8

 

133 / 90

 

74 / 61

 

125 / 129

 

Perhaps the test is only single threaded? But doesn't explain the inconsistency.





PC: R7 3700x 4.3Ghz/32GB DDR4/GTX1070

 


dfnt
1203 posts

Uber Geek

Lifetime subscriber

  #2106383 11-Oct-2018 16:15
Send private message

Ah right that explains it then.

 

I'd almost be inclined to block samknows testing if they didn't bother asking mlab if they could target their servers for the testing


sidefx
3510 posts

Uber Geek

Trusted

  #2106393 11-Oct-2018 16:46
Send private message

dfnt:

 

Ah right that explains it then.

 

I'd almost be inclined to block samknows testing if they didn't bother asking mlab if they could target their servers for the testing

 

 

 

 

Suspect they knew about it since it's listed right here:

 

https://www.measurementlab.net/tests/#current-tests

 

 





"I was born not knowing and have had only a little time to change that here and there."         | Electric Kiwi | Sharesies
              - Richard Feynman


dfnt
1203 posts

Uber Geek

Lifetime subscriber

  #2106404 11-Oct-2018 16:52
Send private message

sidefx:

 

dfnt:

 

Ah right that explains it then.

 

I'd almost be inclined to block samknows testing if they didn't bother asking mlab if they could target their servers for the testing

 

 

 

 

Suspect they knew about it since it's listed right here:

 

https://www.measurementlab.net/tests/#current-tests

 

 

 

 

Probably after the fact, given:

 

It's worth noting we've been hosting M-Lab nodes for around 5 years now, but no one from Comcom or Samknows has spoken to us about the testing so it's all come as a bit of a surprise to us. If you have any interesting information to share, we'd like to know more :)


SiliconAudio
96 posts

Master Geek


  #2106412 11-Oct-2018 17:19
Send private message

Djmixerdomo:

 

hio77:

 

If you google "speedtest" the built in google speedtest does hit the mlab servers.

 

 

That speed testing service is all over the show.

 

I just did five tests and got some interesting results:

 

Down / Up Mbps

 

77 / 92

 

104 / 8

 

133 / 90

 

74 / 61

 

125 / 129

 

Perhaps the test is only single threaded? But doesn't explain the inconsistency.

 

 

Just did a quick test from work and yes, the m-lab results are all over the place.  I'm seeing much more consistent results from speedtest.net even if I don't stick to a specific server.  m-lab tests are often much slower and highly variable.  The speedtest.net results pretty much match our speed on the dot +/- 5 Mbps (Kordia uncontested 100 Mbps).  The m-lab tests showed anything from 48 to 93 Mbps down and 3 to 95 Mbps up (yes, you read that right, 3 Mbps up)!

 

If this is the only domestic service used by SamKnows, I am not surprised that my results are so crap!  Again I say, based on what I am seeing, SamKnows data cannot be trusted.


hio77
'That VDSL Cat'
12626 posts

Uber Geek

Trusted
Subscriber

  #2106413 11-Oct-2018 17:23
Send private message

SiliconAudio:

 

 

 

Just did a quick test from work and yes, the m-lab results are all over the place.  I'm seeing much more consistent results from speedtest.net even if I don't stick to a specific server.  m-lab tests are often much slower and highly variable.  The speedtest.net results pretty much match our speed on the dot +/- 5 Mbps (Kordia uncontested 100 Mbps).  The m-lab tests showed anything from 48 to 93 Mbps down and 3 to 95 Mbps up (yes, you read that right, 3 Mbps up)!

 

If this is the only domestic service used by SamKnows, I am not surprised that my results are so crap!  Again I say, based on what I am seeing, SamKnows data cannot be trusted.

 

 

to be clear, @Djmixerdomo was right. the google test is single threaded.

 

this compares to speedtest.net which is multi-threaded.

 

 

 

I'm not sure the extreme take they must not be trusted is a little far. Sure they appear to be having a slight teething issue.





#include <std_disclaimer>

 

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

 


Talkiet
4485 posts

Uber Geek

Trusted

  #2106417 11-Oct-2018 17:31
Send private message

The data collection is in a test phase at the moment. There will be no RSP comparison reports generated from current data and I suspect they are still making sure the servers are dimensioned appropriately.

 

Throwing around comments like I see here is irresponsible, especially at this stage.

 

All of the issues mentioned here have been hashed out between industry experts and Samknows/Comcom. I think it's fair to say that the technical RSP representatives were very impressed with the Samknows expertise and methodology - both for data collection and for the analysis constraints. It's in a different league to the previous Comcom testing.

 

If you're still seeing really inconsistent results after they launch properly (after you see their PUBLIC reports) then it's reasonable to ask questions about why the data seems inconsistent - until then you don't know if the data is going to be used, or even if it's using final test servers or traffic paths.

 

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.


1 | ... | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19
View this topic in a long page with up to 500 replies per page Create new topic





News »

Vodafone enables 5G roaming - for when international travel comes
Posted 30-Oct-2020 15:03


Spark awards funding to Kiwi businesses in 5G funding initiative
Posted 30-Oct-2020 14:58


Huawei launches IdeaHub Pro in New Zealand
Posted 27-Oct-2020 16:41


Southland-based IT specialist providing virtual services worldwide
Posted 27-Oct-2020 15:55


NASA discovers water on sunlit surface of Moon
Posted 27-Oct-2020 08:30


Huawei introduces new features to Petal Search, Maps and Docs
Posted 26-Oct-2020 18:05


Nokia selected by NASA to build first ever cellular network on the Moon
Posted 21-Oct-2020 08:34


Nanoleaf enhances lighting line with launch of Triangles and Mini Triangles
Posted 17-Oct-2020 20:18


Synology unveils DS16211+
Posted 17-Oct-2020 20:12


Ingram Micro introduces FootfallCam to New Zealand channel
Posted 17-Oct-2020 20:06


Dropbox adopts Virtual First working policy
Posted 17-Oct-2020 19:47


OPPO announces Reno4 Series 5G line-up in NZ
Posted 16-Oct-2020 08:52


Microsoft Highway to a Hundred expands to Asia Pacific
Posted 14-Oct-2020 09:34


Spark turns on 5G in Auckland
Posted 14-Oct-2020 09:29


AMD Launches AMD Ryzen 5000 Series Desktop Processors
Posted 9-Oct-2020 10:13









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.