[Qlog] After a year, a wild qlog draft-02 appears

Robin MARX <robin.marx@uhasselt.be> Thu, 05 November 2020 17:06 UTC

MIME-Version: 1.0
From: Robin MARX <robin.marx@uhasselt.be>
Date: Thu, 05 Nov 2020 18:05:51 +0100
Message-ID: <CAC7UV9aDfdVfRv0Z1EE8eiS6vYd=EZEBo_qZ=qA2B9z81XqgCQ@mail.gmail.com>
To: qlog@ietf.org
Content-Type: multipart/alternative; boundary="0000000000005c729705b35f1c25"
Archived-At: <https://mailarchive.ietf.org/arch/msg/qlog/fOMmoTnvbfVxyj305CCZ8t5YTQc>
Subject: [Qlog] After a year, a wild qlog draft-02 appears
Precedence: list

Hello everyone,

This mailing list has been particularly quiet for the last year, for which
I am sorry.
This however does not mean that the qlog project was halted, on the
contrary.
Many of you and others in the QUIC community have actively contributed to
the project [0],
leading to the new draft-02 versions of the documents [1][2].

This version includes a large amount of updates, which are reflected in the
changelogs on the bottom of each document. The main change is a
backwards-incompatible switch to a more default JSON serialization format,
which makes it easier to also support other serialization formats down the
line (e.g., binary formats like protocol buffers).

This means that it can be non-trivial to adapt an existing draft-01
implementation to draft-02. As such, I will soon create and share a HOWTO
document with concrete examples on how to accomplish this (similar to what
we had for -00 to -01, see [3]). I am also committed to keeping draft-01
support in the qvis toolsuite [4] for a long time, as this is afaik still
the most popular way of visualizing and analyzing qlogs. As such, no one
needs to rush to support -02. Ironically, qvis itself currently does not
fully support draft-02 yet, but will be updated in the coming weeks.

While draft-02 is a substantial move in the "right direction", it is far
from done, and there remain many open issues on the github repository [5].
However, we have come at a state where I feel it is impossible for me to
make much more meaningful progress without broader community involvement.
While many have provided the occasional (and much appreciated) comments,
the core qlog team can be said to consist of mainly two people (me and
Marten Seemann), who often disagree on the "correct" approach.

As such, I hope to gather some input from the QUIC wg group at IETF 109 on
what the next steps are for qlog in terms of where the work is done, who
does it and to what ultimate goal [6].
I assume most of you here are also on the QUIC mailing list, but if not,
comments on this are of course also welcome on the qlog mailing list.

With best regards,
Robin

[0]: https://qlog.edm.uhasselt.be/anrw/
[1]: https://tools.ietf.org/html/draft-marx-qlog-main-schema-02
[2]:
https://tools.ietf.org/html/draft-marx-qlog-event-definitions-quic-h3-02
[3]:
https://github.com/quiclog/internet-drafts/blob/master/implementers/transition-00-to-01.md
[4]: https://qvis.edm.uhasselt.be/
[5]: https://github.com/quiclog/internet-drafts
[6]: https://mailarchive.ietf.org/arch/msg/quic/yv9FFyXItsKK6m-5I5eRE6jnqR8/
--

Robin Marx
PhD researcher - web performance
Expertise centre for Digital Media

T +32(0)11 26 84 79 - GSM +32(0)497 72 86 94

www.uhasselt.be
Universiteit Hasselt - Campus Diepenbeek
Agoralaan Gebouw D - B-3590 Diepenbeek
Kantoor EDM-2.05

[Qlog] After a year, a wild qlog draft-02 appears Robin MARX