Re: [quicwg/base-drafts] Disallow changes of table size after 0-RTT (#2276)

[Martin Thomson <[e-mail]>]

If we allow 0 to become >0, that isn't fundamentally that different to allowing increases. If you chose 0, then 0-RTT probably isn't that effective either. I am inclined to leave no change as the only option if we go this way.

The alternative is to prohibit wrapping until SETTINGS is received, which I agree is manageable, but a little clumsy. For me, this comes down to whether we think that changing max capacity will be that common. It might be easier to say that changes (increases really) can't be made until you get a full handshake.

We see only slightly more than 1% of our TLS 1.3 connections accepting 0-RTT - or even attempting it. It is early days for 0-RTT, but that suggests that there will be plenty of opportunities - even for a server that prefers 0-RTT - to bump the capacity limit.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/2276#issuecomment-453696458