I agree that this would be quite helpful, especially for the use-case for splitting off traffic that supports SNI to a separate alt-svc. Other corner cases with TLS-terminating middleboxes, I can't see why those alt-svc assignments couldn't persist past network changes. On Wed, Jun 3, 2015 at 8:09 PM, Ryan Hamilton <rch@google.com> wrote: > I'd love to see an extension like this! > > On Tue, Jun 2, 2015 at 12:01 AM, Mark Nottingham <mnot@mnot.net> wrote: > >> >> > On 2 Jun 2015, at 2:13 am, Martin Thomson <martin.thomson@gmail.com> >> wrote: >> > >> > On 31 May 2015 at 18:06, Mark Nottingham <mnot@mnot.net> wrote: >> >> This is an interesting discussion to have in concert with #69 >> regarding extensibility; if we make services containing unrecognised >> extensions must-ignore, it would make this sort of thing much chattier; the >> above as an after-the-fact extension would need to be this on the wire: >> >> >> >> Alt-Svc: h2=":443"; ma=3600, h2=":443"; ma=3600; persist >> > >> > Yep, but if we add it now, that concern is less of a problem because >> > servers can send it will a reasonable expectation of it being >> > understood. >> >> Absolutely. But, there's always the next extension… >> >> >> -- >> Mark Nottingham https://www.mnot.net/ >> >> >> >
Received on Thursday, 4 June 2015 04:03:03 UTC