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

[Marten Seemann <[e-mail]>]

In https://github.com/quicwg/base-drafts/pull/2831#pullrequestreview-253535028, @MikeBishop points out that after a Retry, all 0-RTT packets sent in the first flight are still valid, although they will probably have been dropped by the server (unless there's significant reordering). That means that a client SHOULD retransmit all 0-RTT packets in the second flight.

We now have two different scenarios:

1. In the case of a HelloRetryRequest, the client is prohibited from sending 0-RTT, and has to cancel retransmissions for all outstanding 0-RTT packets.
2. In the case of Retry, the 0-RTT packets from the first flight are declared lost (without any effect on the cwnd though) and retransmitted.

Considering that Retry adds a roundtrip anyway (and thereby limits the usefulness of 0-RTT), and that we only expect it to be used by a server that's under attack (which makes it likely to refuse 0-RTT later in the handshake), does it really make sense to send 0-RTT in that case?

-- 
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