Re: [quicwg/base-drafts] Does it make sense to try 0-RTT after Retry? (#2842)

[Marten Seemann <[e-mail]>]

@kazuho, thanks for pointing that out. My point is that we should change the draft (i.e. this is a design issue, not an editorial issue).

> I am not sure if I like this ambiguity (because I think most if not all of us will be dropping 0-RTT packets that carry the original DCID when sending a Retry), but IMO the draft is clear.

The reason a server (or a LB) performs a Retry is that it either

- has reasons to suspect that the client isn't reachable at the 2-tuple it sent the packet from, or
- that it is under load, and either uses the Retry to redirect traffic to a different backend server, or just to buy itself some time

Under both circumstances, a server wouldn't want to allocate additional state for buffering 0-RTT packets.

I'm therefore suggesting to change the draft such that:
1. A server that performs a Retry MUST discard 0-RTT packets.
2. A client MUST NOT use 0-RTT after a Retry was performed.

Does that make sense to you?

-- 
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/2842#issuecomment-505249709