User comments on ISPs
  >> Virgin Media (Cable)


Register (or login) on our website and you will not see this ad.


Pages in this thread: 1 | [2] | 3 | 4 | 5 | (show all)   Print Thread
Standard User Chrysalis
(legend) Sun 14-May-17 18:20:15
Print Post

Re: Bufferbloat?


[re: MrSaffron] [link to this post]
 
and with 100 download threads smile

Sky Fibre Pro BQM - IPv4 BQM - IPv6
Standard User Daemon66
(member) Mon 15-May-17 09:51:29
Print Post

Re: Bufferbloat?


[re: JohnR] [link to this post]
 
Testing only when there is a problem is an obvious issue that I'm sure TB adjust for in their statistics, don't they?
Personnally I prefer to test before a problem becomes apparent in the hope that I can avert it. It is a shame the test becomes meaningless if it reports a connection that works well as being bad. I extensively investigated the connection in a fruitless effort to identify the issue TB told me I had only to find there is no issue.
Standard User Daemon66
(member) Mon 15-May-17 10:00:41
Print Post

Re: Bufferbloat?


[re: Chrysalis] [link to this post]
 
In reply to a post by Chrysalis:
your reasoning is a bit like saying speedtests are unfair on dialup tech smile
Sorry, I don't get your meaning. Speedtests reflect the transfer speed of the connection whatever the technology. Though some test do still attempt to highlight issues with some technologies more than others by focusing on single-threaded tests that are more sensitive to buffering (regardless of what some may assert, TCP is very sensitive to buffering).
It is well known that cable buffers more than other technologies, this is why so many complain about the level of jitter on cable. This buffering means that TB now show my cable connection as bad quality.


Register (or login) on our website and you will not see this ad.

Administrator MrSaffron
(staff) Mon 15-May-17 12:40:22
Print Post

Re: Bufferbloat?


[re: Daemon66] [link to this post]
 
Can you show me where we said your connection is bad quality?

Our explainer says....
Bufferbloat is a measure of how delayed packets are due to excessive buffering in broadband routers and other locations on the Internet. A is good, F is bad and is only something to be worried about if you upload a lot while also downloading or vice versa. Enabling QoS on some routers can improve performance.

And if as asserted the bufferbloat testing is weighted against cable why would the test at http://tbb.st/1494845036839948755
have shown with grades B and A.

As for single thread - nothing to do with buffer bloat at all, since we don't do any latency testing during the single thread test.

http://tbb.st/1494844753746210455

If you are worried so much about bufferbloat and single thread then perhaps http://www.broadband.co.uk/broadband-speed-test/ is a better test for you to use.

The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User Daemon66
(member) Mon 15-May-17 14:11:14
Print Post

Re: Bufferbloat?


[re: MrSaffron] [link to this post]
 
This is like talking to a politician, never a straight answer and often answers a different question:
In reply to a post by MrSaffron:
Can you show me where we said your connection is bad quality?
Other than what I said in my original post? Because you don't actually allow me to see/share the quality results after the test?
In reply to a post by MrSaffron:
And if as asserted the bufferbloat testing is weighted against cable why would the test at http://tbb.st/1494845036839948755
have shown with grades B and A.
Did it? I can't see any ratings for bufferbloat on that link. Interesting that that test doesn't get near the full upload speed though.
In reply to a post by MrSaffron:
As for single thread - nothing to do with buffer bloat at all, since we don't do any latency testing during the single thread test.
For more information on the effect of buffering on TCP throughput here is a simple primer: https://en.wikipedia.org/wiki/TCP_tuning
In reply to a post by MrSaffron:
If you are worried so much about bufferbloat and single thread then perhaps http://www.broadband.co.uk/broadband-speed-test/ is a better test for you to use.
That's not the point, I'm not worried about it, it has no effect on my connection, my connection is not subject to bufferbloat. I just want to know why you are saying it is and hence reporting my connection as bad quality.
Administrator MrSaffron
(staff) Mon 15-May-17 14:23:55
Print Post

Re: Bufferbloat?


[re: Daemon66] [link to this post]
 
We said the buffer bloat was poor not the WHOLE connection was bad quality, I cannot look at the actual numbers for your connection as you've not linked to a test.

And the primer tells me little I did not know - but reminds me to highlight that large buffers are ideally used on high latency networks, which I am sure we all agree is not what Virgin Media is.

As for the test linked to, what speed do you expect the upload to be on a 200 Mbps service?
The service is advertised currently as up to 12 Mbps, but have not heard any announcement to say Virgin Media has upgraded all 200 Mbps service customers to the latest 12 Mbps upload.

Run another test, and link to it so I can look at the underlying numbers, for all I know there may be a mathematical error, but without an example of test showing what you say its hard to say WHY we are saying buffer bloat appears to exist on that connection.

The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User Daemon66
(member) Mon 15-May-17 15:54:47
Print Post

Re: Bufferbloat?


[re: MrSaffron] [link to this post]
 
Here is the test to which I referred, but the link doesn't show the bufferbloat/quality rating: https://www.thinkbroadband.com/speedtest/14944862896...

NOTE: That I didn't turn off my Nest Cam during the test so upload is a little lower than it could be, probably by about 0.5Mbps.
Administrator MrSaffron
(staff) Mon 15-May-17 16:35:31
Print Post

Re: Bufferbloat?


[re: Daemon66] [link to this post]
 
Correct the link does not show the buffer bloat figures, and probably not going to.

So we recorded 18ms average TCP latency
When downloading latency was measured at 80ms
When uploading latency was measured at 131ms

So download was just in the D zone, and during upload in the F zone.

Maybe there is a wider debate to be had on whether to qualify for grade A the latency needs to be less than double. The original grading was determined by looking at the spread of values across the ISP range.

Existing plans were underway to make more info available at the end of a test to show people these values, so they can ignore the grading system and decide for themselves is a jump from 18ms to 131ms is an issue.

The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User Daemon66
(member) Tue 16-May-17 09:03:05
Print Post

Re: Bufferbloat?


[re: MrSaffron] [link to this post]
 
Finally, thanks for the explanation.

Since anything < 150ms is generally considered perfectly fine for VOIP and < 100 ms is fine for LoL (not great but fine). I guess that's why I don't see any issue with my connection, though of course you appear to have a different view.

I would say that the grading should reflect the likelihood of bufferbloat actually causing an issue.
Administrator MrSaffron
(staff) Tue 16-May-17 09:36:31
Print Post

Re: Bufferbloat?


[re: Daemon66] [link to this post]
 
So lets make this a RFC call and hear what people want for the grades?

i.e. latency at idle and what multiplier to consider acceptable for each grade?

The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Pages in this thread: 1 | [2] | 3 | 4 | 5 | (show all)   Print Thread

Jump to