Re: [quicwg/base-drafts] Desirable behavior when it takes time to derive the traffic keys for the next PN space (#3821)

[Martin Thomson <[e-mail]>]

We encountered a problem similar to the one described here in our stack.  The latter part of a coalesced packet was being dropped because we had to spend time validating the certificate.  We ultimately added a small amount of buffering to address this.  And I have debugged a case where handshakes were failing due to the interaction between PTO timers in different packet number spaces.  For that, I don't have a good solution yet; I'm not convinced that deferring PTO is the right answer.

But I don't think that this requires anything in specifications, aside from perhaps a caution, noting that we already have a caution in [Section 5.7 of the TLS draft](https://quicwg.org/base-drafts/draft-ietf-quic-tls.html#pre-hs-protected).  That already suggests and permits (though doesn't recommend) buffering.

We could add a note that failing to buffer legitimate packets could lead to marking delivered packets as lost unnecessarily and performance degradation

-- 
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/3821#issuecomment-654012206