Re: [Cerowrt-devel] archer c7 v2, policing, hostapd, test openwrt build

[Sebastian Moeller <moeller0@gmx.de>]

[-- Attachment #1: Type: text/plain, Size: 317 bytes --]

Hi Jonathan, hi List,


So I got around to a bit of rrul testing of the dual egress idea to asses the cost of IFB, but the results are complicated (so most likely I screwed up). On an wndr3700v2 on a 100Mbps/40Mbps link I get the following (excuse the two images either the plot is intelligble or the legend...):

[-- Attachment #2: dual_egress_test_20150324_legend.eps --]
[-- Type: image/eps, Size: 133684 bytes --]

[-- Attachment #3: dual_egress_test_20150324.eps --]
[-- Type: image/eps, Size: 106766 bytes --]

[-- Attachment #4: Type: text/plain, Size: 1889 bytes --]



Only in case of shaping the total bandwidth to the ~70Mbps this router can barely do can I see an effect of the dual egress instead of the IFB based ingress shaper. So column 7 (ipv4) and column 8 (ipv6) are larger than columns 9 (ipv4) and 10 (ipv6) showing that dual egress instead of egress and ingress effective upload increases by < 10 Mbps (while download and latency stay unaffected). That is not bad, but also does not look like the IFB is the cost driver in sqm-scripts, or does it? Also as a corollary of the data I would say, my old interpretation that we hit a limit at ~70Mbps combined traffic might not be correct in that ingress and egress might carry slightly different costs, but then thins difference is not going to make a wndr punch way above its weight…

Best Regards
	Sebastian 




On Mar 23, 2015, at 17:09 , Sebastian Moeller <moeller0@gmx.de> wrote:

> Hi Jonathan,
> 
> On Mar 23, 2015, at 14:43 , Jonathan Morton <chromatix99@gmail.com> wrote:
> 
>> 
>>> On 23 Mar, 2015, at 08:09, Sebastian Moeller <moeller0@gmx.de> wrote:
>>> 
>>> It obviously degrade local performance of se00 and hence be not a true solution unless one is happy to fully dedicate a box as shaper ;)
>> 
>> Dedicating a box as a router/shaper isn’t so much of a problem, but shaping traffic between wired and wireless - and sharing the incoming WAN bandwidth between them, too - is.  outer
> 
> 	Exactly the sentiment I had, but less terse and actually understandable ;)
> 
>> It’s a valid test, though, for this particular purpose.
> 
> 	Once I get around to test it, I should b able to share some numbers…
> 
> Best Regards
> 	Sebastian
> 
>> 
>> - Jonathan Morton
>> 
> 
> _______________________________________________
> Cerowrt-devel mailing list
> Cerowrt-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cerowrt-devel