W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2020

Re: Priority signals in HTTP/2

From: Martin Thomson <mt@lowentropy.net>
Date: Mon, 21 Dec 2020 10:09:22 +1100
Message-Id: <ea1bea44-0536-484f-a0c1-548f3644ded4@www.fastmail.com>
To: "Willy Tarreau" <w@1wt.eu>
Cc: ietf-http-wg@w3.org
Hi Willy,

On Sun, Dec 20, 2020, at 04:04, Willy Tarreau wrote:
> I think I'm OK with this plan. I'm just thinking, do we have an exhaustive
> list of the most annoying problems caused by the current scheme and do we
> think that we could bring back some value if they were at least partially
> addressed ? If so, maybe we could have the following recommendations by
> order of preference:
>   - new priority scheme (yet to be defined)
>   - "fixed" version of the previous one to be used as a second option,
>     expected to work better than nothing with existing implementations;
>   - no priority handling at all (i.e. round robin or random)

I don't want to get too much into the shortcomings of the existing scheme, except for those things we know to be firmly true.  The only thing that I think can be said definitively is that attempting to send responses one at a time is better than trying to evenly share available capacity between multiple responses.

I expect that the new thing we ship will be much better than that, but it's probably best to let that specification speak for itself.
Received on Sunday, 20 December 2020 23:09:55 UTC

This archive was generated by hypermail 2.4.0 : Sunday, 20 December 2020 23:09:56 UTC