Cake - FQ_codel the next generation
 help / color / mirror / Atom feed
From: Dave Taht <dave.taht@gmail.com>
To: Jordan Szuch <jordan@inacomptc.com>
Cc: Cake List <cake@lists.bufferbloat.net>
Subject: Re: [Cake] pcq on mikrotik
Date: Thu, 21 Jul 2022 15:01:49 -0700	[thread overview]
Message-ID: <CAA93jw78j9r+Q=8=_uBqRBiWC5piisXHBKbkcatUBWpa_uJh4g@mail.gmail.com> (raw)
In-Reply-To: <005101d89d47$ad0870c0$07195240$@inacomptc.com>

[-- Attachment #1: Type: text/plain, Size: 3518 bytes --]

Thank you jordan!

That is a really remarkable difference in performance between these
two settings. You are only getting a peak of 40Mbit download out of
pcq, vis about 100 on cake. Cake is winning across the board, here...
That said, cake is doing synchronized drops on roughly a 10s interval.
I really wish (can you file a feature request?) that we could do
inbound shaping using cake's default shaper.

A couple other notes below:

On Thu, Jul 21, 2022 at 2:20 PM <jordan@inacomptc.com> wrote:
>
> Hey Dave,
>
> I have a Mikrotik setup for testing and running cake right now. Figured I could run some tests through it. I plugged in the PCQ queues Mikrotik has in there by default and ran RRUL for both queue configurations while waiting on some work to finish. If you'd like some different tests run through it let me know and I'll give it a go!
>
> Here's the settings for Cake and PCQ
>
> add cake-ack-filter=filter cake-diffserv=diffserv4 cake-memlimit=32.0MiB \
>     cake-mpu=64 cake-nat=yes cake-overhead=18 cake-overhead-scheme=docsis \
>     kind=cake name=cake-up
> add cake-diffserv=diffserv4 cake-memlimit=32.0MiB cake-mpu=64 cake-nat=yes \
>     cake-overhead=18 cake-overhead-scheme=docsis cake-wash=yes kind=cake \
>     name=cake-down
>
>     name="pcq-upload-default" kind=pcq pcq-rate=0 pcq-limit=50KiB pcq-classifier=src-address pcq-total-limit=2000KiB
>      pcq-burst-rate=0 pcq-burst-threshold=0 pcq-burst-time=10s pcq-src-address-mask=32 pcq-dst-address-mask=32

I note I only plowed through the pcq docs this morning, so my thoughts
are less than accurate.

pcq-burst-time=10s is calculating an average over 10s to determine how
to shape. even 1s is a lot, IMHO.

Similarly the classifiers should be unneeded with cake.

>      pcq-src-address6-mask=128 pcq-dst-address6-mask=128
>
>         name="pcq-download-default" kind=pcq pcq-rate=0 pcq-limit=50KiB pcq-classifier=dst-address pcq-total-limit=2000KiB

not clear what the pcq-limit or pcq-totale-limit does

>      pcq-burst-rate=0 pcq-burst-threshold=0 pcq-burst-time=10s pcq-src-address-mask=32 pcq-dst-address-mask=32
>      pcq-src-address6-mask=128 pcq-dst-address6-mask=128
>
> /queue simple
> add bucket-size=0.001/0.001 dst=ether1 max-limit=10M/100M name=\
>     "Spectrum Cake3" queue=pcq-upload-default/pcq-download-default target=\
>     bridge

I am under the impression that such a small bucketsize is problematic. .01?

>
>     or
>
>     add bucket-size=0.001/0.001 dst=ether1 max-limit=10M/100M name=\
>     "Spectrum Cake3" queue=cake-up/cake-down target=\
>     bridge
>
> And here's a link to the flent data for you or anyone else to take a look at
>
> https://1drv.ms/u/s!Ap4u4Rte63FqjrkzFZJap8UuuSaraQ?e=zWG7xc
>
>
> -----Original Message-----
> From: Cake <cake-bounces@lists.bufferbloat.net> On Behalf Of Dave Taht via Cake
> Sent: Thursday, July 21, 2022 2:41 PM
> To: Cake List <cake@lists.bufferbloat.net>
> Subject: [Cake] pcq on mikrotik
>
> at some point I'd like to compare cake vs:
>
> https://wiki.mikrotik.com/wiki/Manual:Queues_-_PCQ
>
> --
> FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
> Dave Täht CEO, TekLibre, LLC
> _______________________________________________
> Cake mailing list
> Cake@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cake
>


-- 
FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC

[-- Attachment #2: compared.png --]
[-- Type: image/png, Size: 89802 bytes --]

[-- Attachment #3: rrul_-_pcq-spectrum-fqcint-rb5009-rrul-300.png --]
[-- Type: image/png, Size: 353657 bytes --]

      reply	other threads:[~2022-07-21 22:02 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-21 18:40 Dave Taht
2022-07-21 21:20 ` jordan
2022-07-21 22:01   ` Dave Taht [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://lists.bufferbloat.net/postorius/lists/cake.lists.bufferbloat.net/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAA93jw78j9r+Q=8=_uBqRBiWC5piisXHBKbkcatUBWpa_uJh4g@mail.gmail.com' \
    --to=dave.taht@gmail.com \
    --cc=cake@lists.bufferbloat.net \
    --cc=jordan@inacomptc.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox