RE: Benjamin Kaduk's Discuss on draft-ietf-httpbis-http2bis-06: (with DISCUSS and COMMENT)

I don't think I was mentioned three times, but I shall consider myself summoned nonetheless.



The H3 text is:

Clients SHOULD NOT open more than one HTTP/3 connection to a given IP address and UDP port, where the IP address and port might be derived from a URI, a selected alternative service ([ALTSVC]), a configured proxy, or name resolution of any of these. A client MAY open multiple HTTP/3 connections to the same IP address and UDP port using different transport or TLS configurations but SHOULD avoid creating multiple connections with the same configuration.



The HTTP/3 text does allow for differences of TLS configurations, which would encompass different certificates and/or SNI values, so I don’t think there’s necessarily a conflict here.



The H3 text was written in contemplation of Alt-Svc being the primary means of finding the endpoint and the Alt-Svc entries potentially covering a different set of hosts than the endpoint has certificates for.  By keying on IP address, it means that two Alt-Svc advertisements pointing to different hostnames which resolve to the same IP address would still enable connection reuse if the certificate covers both origins, while the HTTP/2 text would advise separate connections to that IP address for the two origins.



I think we could leave the text as-is in both documents and be fine.  I think we could normalize the documents to either version and be fine.  I tend to prefer more coalescing than less, FWIW.



-----Original Message-----
From: Martin Thomson <mt@lowentropy.net>
Sent: Friday, January 7, 2022 9:44 AM
To: Benjamin Kaduk <kaduk@mit.edu>; The IESG <iesg@ietf.org>
Cc: draft-ietf-httpbis-http2bis@ietf.org; httpbis-chairs@ietf.org; ietf-http-wg@w3.org; Mark Nottingham <mnot@mnot.net>
Subject: Re: Benjamin Kaduk's Discuss on draft-ietf-httpbis-http2bis-06: (with DISCUSS and COMMENT)



> Section 9.1

>

>    Clients SHOULD NOT open more than one HTTP/2 connection to a given

>    host and port pair, where the host is derived from a URI, a selected

>    alternative service [ALT-SVC], or a configured proxy.

>

> quic-http has similar text (in §3.3), but it refers to a given IP

> address and port, rather than host and port.  Is the difference

> between host and IP address significant when comparing h/2 and h/3?

> (When using IP addresses, we of course have to additionally talk about

> name resolution of the other types of identifier.)



I honestly don't know.  I think perhaps host is better in this case in the sense that clients aim to connect to hosts and connection coalescing is not a requirement, just permitted (as noted in the text that follows).  I'm not sure that it really matters ultimately, but it's worth checking.



Perhaps Mike Bishop can help us here.