Re: [quicwg/base-drafts] Rework Key Update (#2237)

[Martin Thomson <[e-mail]>]

At this stage, I'm going to put this on the agenda for discussion in Tokyo.  I think that we have three valid options, though we might have ruled the first out:

1. use acknowledgements to regulate multiple updates
2. use two phase bits, representing a cyclic sequence (0, 1, 2, 3) and allow two updates ahead
3. use one phase bit and a second bit indicating whether key updates are permitted

1 is theoretically possible, but likely to be impractical in many implementations, so we can rule that out.

This PR is currently 2, though I'm biased slightly toward 3 now.  It doesn't have an error condition where the phase is 3 ahead of expected, but the packet number is higher.  More importantly, it reduces the number of active keys after the handshake is complete to at most 2.  The drawback is that you can block key updates.  That said, some people might consider that to be a feature.

-- 
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/pull/2237#issuecomment-450593563