Clarification on changes

frochet / draft-rochet-reverso-for-quic

0 stars 0 forks source link

Clarification on changes #1

Open frochet opened 4 weeks ago

frochet commented 4 weeks ago

Marten Seemann on the mailing list:

Let me elaborate on why I think that these are drastic changes to QUIC. Reversing the wire representation of frames is one thing, but adding a number of fields to the Short Header, and mandating that any packet can only carry a single STREAM frame, seems to be a much larger change, requiring non-trivial changes to the packetization logic.

We should probably clarify sooner than Section "Packing Frames" that there is no mandate to carry a single Stream Frame.

frochet commented 3 weeks ago

Interesting discussion on the quicdev slack resulted into the following:

~~No implementation seems to currently doing Stream frame multiplexing~~ (edit)
- quinn & s2n seem doing it in some cases
Reverso gives more incentives to avoid stream frame multiplexing, but does not prevent to do it
Instead of recommending to avoid multiplexing, we should simply state the facts (zerocopy for the first one, a copy for the others), and that it may exists use cases for multiplexing.