Re: 0-RTT Design for HTTP/2

On Fri, Dec 18, 2020, at 04:52, David Schinazi wrote:
> Could you please elaborate on your statement "My view is that we can't 
> really deploy ALPS in an incremental fashion for HTTP/2.  Part of what 
> I'm looking for here is an incremental solution."? I'm not seeing what 
> makes your proposal more incremental than ALPS, so perhaps I'm missing 
> something here.

I gave a longer response regarding suitability of ALPS in that thread. As did Cory.

I can see how you might choose to opportunistically use ALPS, but I don't think that you can add it trivially to HTTP/2 implementations.  It's more of a case of not wanting it than it being functionally impossible; we could likely make anything work if we assume good integration of stack layers, but that isn't how HTTP/2 has been deployed and retrofitting something in that way would only result in a privileged few having the capability.  I would rather see it deployed for new protocols than try to have its presence alter how this existing protocol works.

Received on Thursday, 17 December 2020 22:13:23 UTC