- From: Willy Tarreau <w@1wt.eu>
- Date: Mon, 7 Feb 2022 06:44:42 +0100
- To: Martin Thomson <mt@lowentropy.net>
- Cc: Wenbo Zhu <wenboz@google.com>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>, Tommy Pauly <tpauly@apple.com>, Christopher Wood <caw@heapingbits.net>
On Mon, Feb 07, 2022 at 09:08:04AM +1100, Martin Thomson wrote: > On Sun, Feb 6, 2022, at 04:10, Wenbo Zhu wrote: > > I wonder if an http-wg version of "ASN.1" definition is needed here. We > > certainly won't switch to bhttp anytime soon, and there will likely be > > performance gaps too to adopt such a new encoding. > > That's a strong claim. Is there something in the design that suggests to you > a performance deficit? I agree with Martin here, my understanding is that it requires the same job as translating H1 to H2, without the cost of looking up fields in the HPACK table nor encoding the output. This looks like a single-stream, simplified H2, and as such I do not see any particular performance issue here. Willy
Received on Monday, 7 February 2022 05:45:01 UTC