Many ISPs need the kinds of quality shaping cake can do
 help / color / mirror / Atom feed
From: Dave Taht <dave.taht@gmail.com>
To: Jan Rovner <jan.rovner@diadema.cz>
Cc: "libreqos@lists.bufferbloat.net" <libreqos@lists.bufferbloat.net>
Subject: Re: [LibreQoS] A newbie question - bridge mode
Date: Thu, 29 Dec 2022 08:34:15 -0800	[thread overview]
Message-ID: <CAA93jw5OqDSfLh60KBumTWLmG8VhPFsUrK26Ej8LTehU_H7Yzw@mail.gmail.com> (raw)
In-Reply-To: <f1a596d72f2c46e4b9e51e836a542be4@diadema.cz>

On Thu, Dec 29, 2022 at 8:15 AM Jan Rovner via LibreQoS
<libreqos@lists.bufferbloat.net> wrote:
>
> Hello,
>
> I like libreqos project and initially I would like to ask about the possibility of operating in bridge mode, even not oficially fully  supported.
>
> My question targets underlying technologies like XDP a others and their performance: is it possible to achieve the same (or similar) performance on QoS on bridging traffic shaper - with VLANs - as on routed setup?

Yes and no. Our biggest testbed is setup as a "router on a stick",
with two vlans, bridging via xdp. As soon as we put routing into the
equation a whole lot more of the stack has to get involved, and
performance suffers. (currently cracking 25Gbit in bridged mode, and
we might get to 50Gbit soon).

I don't have good numbers for what a routed setup would look like, I
suspect they would be below 10Gbit on 16 core hardware (and we
wouldn't be able to scale to use those cores very efficiently). Yes!
Doing this stuff in a routed setup is very feasible, just not
something we're exploring actively right now. Much of the the code can
be made to work routed fairly easily.  A hope I have is a network card
will show up that can do LPM or some set of more intelligent offloads
- as the cake and fq_codel algorithms are very lightweight compared to
the cost of reading data from the network, routing it, HTB or LPM.

Most of the work these days is taking place on matrix -
#libreqos:matrix.org, please feel free to join us there.

Also, a demo of v1.4 WIP can be found here: https://payne.taht.net/

Thanks for your interest!

> Thank you.
> Jan Rovner
>
>
> _______________________________________________
> LibreQoS mailing list
> LibreQoS@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/libreqos



-- 
This song goes out to all the folk that thought Stadia would work:
https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-6981366665607352320-FXtz
Dave Täht CEO, TekLibre, LLC

  parent reply	other threads:[~2022-12-29 16:34 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-12-29 16:15 Jan Rovner
2022-12-29 16:27 ` Herbert Wolverson
2022-12-29 16:34 ` Dave Taht [this message]
2023-03-18 19:16 ` [LibreQoS] Bifrost XDP-Accelerated Bridge + preserve VLAN tags? Jan Rovner

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/libreqos.lists.bufferbloat.net/

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

  git send-email \
    --in-reply-to=CAA93jw5OqDSfLh60KBumTWLmG8VhPFsUrK26Ej8LTehU_H7Yzw@mail.gmail.com \
    --to=dave.taht@gmail.com \
    --cc=jan.rovner@diadema.cz \
    --cc=libreqos@lists.bufferbloat.net \
    /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