Re: [quicwg/base-drafts] Use a "stream" for transmitting CIDs (#1826)

[Kazuho Oku <[e-mail]>]

@MikeBishop FWIW, I do not think that flow control of CIDs is essential, even though I included that in this proposal. CRYPTO streams do not have flow control, and "CID frames" would work fine without flow control too (assuming that we continue allowing the receiver to "forget" CIDs).

> there's also the retire issue -- stream data is never really "retired." 

Would you mind elaborating on this? For the CID, the proposal has a RETIRE message. But I assume you are referring to something else.

> Also, while there are reasons we want all sequence numbers to be generated in order, there's no reason you need to block using CID 7 because you haven't seen CID 6.

That's correct. But it is still a requirement for book keeping. My point is that it would be worthwhile to consider reusing the "streams" approach, if we are to require book keeping using sequence numbers in a situation where HoLB is not a concern.

-- 
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/1826#issuecomment-426430637