Hi Mohit,

Thanks for doing the comparison!

As for "any other material related to this?" I've had a dig around, and there is not a lot. Rasool (the FreeBSD fq-pie author) did write up a university tech report in April 2016, "Dummynet AQM v0.2 – CoDel, FQ-CoDel, PIE and FQ-PIE for FreeBSD’s ipfw/dummynet framework" (wayback machine link). There's also a small mention in an LCN 2017 paper titled "Characterising LEDBAT Performance Through Bottlenecks Using PIE, FQ-CoDel and FQ-PIE Active Queue Management" (ieee link, author preprint). Hope these are of some (historical) interest to your I-D (and the wider bufferbloat mailing list, too). I'm unaware of any production deployment of FreeBSD's fq-pie.

cheers,
gja

On 2025-03-18 18:34, Mohit P. Tahiliani wrote:
Hi Grenville,

Thanks for the email and the commit link!

Sorry for the late reply. I was traveling, and also going through the code which you shared.

I believe the implementation of FQ-PIE in FreeBSD matches the proposed ID very closely. I checked the code for the following:

1. There is a per queue PIE instance
2. Implementation of ECN support as described in Section 5.1 of RFC 8033 (https://datatracker.ietf.org/doc/html/rfc8033#section-5.1)
3. Use timestamps as a default approach to calculate queue delay instead of Little's Law
4. The functionality of FQ to be similar to the one used in FQ-CoDel

and I found that all of the above are the same as mentioned in the ID. I'll edit the ID and mention this implementation. As of now, the ID mentions two implementations: Linux and ns-3.

Do you have any measurements performed with this implementation? or any other material related to this? I see it mentions that this is a beta implementation, but in case it was used in any real deployment or tested in live networks?

Thanks and Regards,
Mohit P. Tahiliani

On Tue, Mar 11, 2025 at 6:51 AM grenville armitage via Bloat <bloat@lists.bufferbloat.net> wrote:
Hi Mohit,

I wonder what you think of the fq-pie developed for FreeBSD back in 2015 and merged in May 2016? (commit here.)  It has been too long since I last looked, so would love your thoughts on how it aligns (or doesn't) with https://datatracker.ietf.org/doc/draft-tahiliani-tsvwg-fq-pie/?

cheers,
gja

On 2025-03-11 06:01, Mohit P. Tahiliani via Bloat wrote:
Hi all,

I have been working on an Internet Draft on Flow Queue PIE (FQ-PIE) [https://datatracker.ietf.org/doc/draft-tahiliani-tsvwg-fq-pie/], and recently submitted a revised version.

It has been incorporated into the mainline Linux kernel as a queuing discipline (qdisc). Additionally, an implementation of FQ-PIE is available in the ns-3 network simulator.

If you have any suggestions or feedback, please let me know.

Thanks and Regards,
Mohit P. Tahiliani

---------- Forwarded message ---------
From: <internet-drafts@ietf.org>
Date: Tue, Mar 4, 2025 at 12:56 AM
Subject: New Version Notification for draft-tahiliani-tsvwg-fq-pie-01.txt
To: Mohit P. Tahiliani <tahiliani@nitk.edu.in>


A new version of Internet-Draft draft-tahiliani-tsvwg-fq-pie-01.txt has been
successfully submitted by Mohit Tahiliani and posted to the
IETF repository.

Name:     draft-tahiliani-tsvwg-fq-pie
Revision: 01
Title:    Flow Queue PIE: A Hybrid Packet Scheduler and Active Queue Management Algorithm
Date:     2025-03-03
Group:    Individual Submission
Pages:    8
URL:      https://www.ietf.org/archive/id/draft-tahiliani-tsvwg-fq-pie-01.txt
Status:   https://datatracker.ietf.org/doc/draft-tahiliani-tsvwg-fq-pie/
HTML:     https://www.ietf.org/archive/id/draft-tahiliani-tsvwg-fq-pie-01.html
HTMLized: https://datatracker.ietf.org/doc/html/draft-tahiliani-tsvwg-fq-pie
Diff:     https://author-tools.ietf.org/iddiff?url2=draft-tahiliani-tsvwg-fq-pie-01

Abstract:

   This document presents Flow Queue Proportional Integral controller
   Enhanced (FQ-PIE), a hybrid packet scheduler and Active Queue
   Management (AQM) algorithm to isolate flows and tackle the problem of
   bufferbloat.  FQ-PIE uses hashing to classify incoming packets into
   different queues and provide flow isolation.  Packets are dequeued by
   using a variant of the round robin scheduler.  Each such flow is
   managed by the PIE algorithm to maintain high link utilization while
   controlling the queue delay to a target value.



The IETF Secretariat



_______________________________________________
Bloat mailing list
Bloat@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/bloat

_______________________________________________
Bloat mailing list
Bloat@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/bloat