[Cake] WAN ingress rate with concurrent downloads

[Noah Causin]

I notice this behavior as well.  I believe it is the fact that SQM is 
not the true bottleneck.  Senders continue to increase their transfer 
rate until active queue management on your router causes them to slow down.

If I watch LEDE's built-in realtime traffic graphs, I do see incoming 
rates higher than the set limit.


On 10/27/2016 2:41 AM, Sebastian Moeller wrote:
> Could you set the ingress shaper to 50% of you true ingress rate and 
> try again, please?
> Also could you post the results of 'tc -d qdisc' with your current 
> settings and with the 50% settings (I will probably ask for more 
> data...)? And finally I believe you are on an ADSL link, what are the 
> sync rates reported by the modem, and which ISP are you with (I ask, 
> as some ISPs actually employ a shaper at their BRAS, so the modem sync 
> rate might not be the actual bottleneck bandwidth). That said, with 
> wrong shaper settings I would rather expect increased latency under 
> load, but not bandwidth violations.
>
> Best Regards
> Sebastian
>
> On October 27, 2016 2:16:06 AM GMT+02:00, "G. Amanakis" 
> <g_amanakis at yahoo.com> wrote:
>
>     I mean that the ingress limit on WAN (or egress limit on LAN)
>     seems to be ignored, and the rate bmon and bwm-ng report is the
>     maximum achievable. I.e. on 450kbyte/s ingress with the limit set
>     at 400kbyte/s, when bittorrent is in use bmon reports 450kbyte/s
>     after 10-20 seconds.
>
>     ------------------------------------------------------------------------
>
>     Cake mailing list
>     Cake at lists.bufferbloat.net
>     https://lists.bufferbloat.net/listinfo/cake
>
> _______________________________________________
> Cake mailing list
> Cake at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cake
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/cake/attachments/20161027/8b25c526/attachment.html>