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
85 posts

Master Geek
+1 received by user: 8


  Reply # 2104031 9-Oct-2018 17:14
Send private message quote this post

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.


Meow
7913 posts

Uber Geek
+1 received by user: 3935

Moderator
Trusted
Lifetime subscriber

  Reply # 2104050 9-Oct-2018 17:59
One person supports this post
Send private message quote this post

@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).





85 posts

Master Geek
+1 received by user: 8


  Reply # 2104053 9-Oct-2018 18:11
Send private message quote this post

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.


3194 posts

Uber Geek
+1 received by user: 911

Trusted

  Reply # 2105672 10-Oct-2018 15:44
Send private message quote this post

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. 


85 posts

Master Geek
+1 received by user: 8


  Reply # 2106323 11-Oct-2018 15:31
Send private message quote this post

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.


'That VDSL Cat'
8706 posts

Uber Geek
+1 received by user: 1886

Trusted
Spark
Subscriber

  Reply # 2106334 11-Oct-2018 15:50
Send private message quote this post

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.


defiant
611 posts

Ultimate Geek
+1 received by user: 295

Lifetime subscriber

  Reply # 2106345 11-Oct-2018 15:57
Send private message quote this post

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..


'That VDSL Cat'
8706 posts

Uber Geek
+1 received by user: 1886

Trusted
Spark
Subscriber

  Reply # 2106376 11-Oct-2018 16:02
Send private message quote this post

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.


224 posts

Master Geek
+1 received by user: 51

Subscriber

  Reply # 2106382 11-Oct-2018 16:13
Send private message quote this post

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: i7 6700k 4.8Ghz/32GB DDR4/GTX1070

 

Server: Dual L5640/36GB DDR3/GTX1060

 

HTPC: R5 2400G/16GB DDR4


defiant
611 posts

Ultimate Geek
+1 received by user: 295

Lifetime subscriber

  Reply # 2106383 11-Oct-2018 16:15
Send private message quote this post

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


3194 posts

Uber Geek
+1 received by user: 911

Trusted

  Reply # 2106393 11-Oct-2018 16:46
Send private message quote this post

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

 

 


defiant
611 posts

Ultimate Geek
+1 received by user: 295

Lifetime subscriber

  Reply # 2106404 11-Oct-2018 16:52
Send private message quote this post

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 :)


85 posts

Master Geek
+1 received by user: 8


  Reply # 2106412 11-Oct-2018 17:19
Send private message quote this post

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.


'That VDSL Cat'
8706 posts

Uber Geek
+1 received by user: 1886

Trusted
Spark
Subscriber

  Reply # 2106413 11-Oct-2018 17:23
Send private message quote this post

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.


3713 posts

Uber Geek
+1 received by user: 2253

Trusted
Spark NZ

  Reply # 2106417 11-Oct-2018 17:31
2 people support this post
Send private message quote this post

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

 

 


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

Twitter »

Follow us to receive Twitter updates when new discussions are posted in our forums:



Follow us to receive Twitter updates when news items and blogs are posted in our frontpage:



Follow us to receive Twitter updates when tech item prices are listed in our price comparison site:



Geekzone Live »

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



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.