Re: [quicwg/base-drafts] Initials carrying different first ClientHello are considered as belonging to different connections (#2076)

[MikkelFJ <[e-mail]>]

Maybe this is a good approach but I'm not sure the exact approach is ideal.

If possible we should aim for a design that is not tightly coupled to TLS crypto content. We could for example add a nonce to the Hello and define that initial destination ID to be a specific well-defined hash of the Hello. The server then only needs to track the initial DCID but can verify incoming packets based on the hash. It is almost the same, but it gives the initial DCID more meaning.

If QUIC v.x does not use TLS, that version needs to make sure the initial DCID matches the critical signature of the first packet, which is then defined differently, but the mechanism remains the same. It might also be helpful to middleboxes?

-- 
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/2076#issuecomment-443311377