Re: [hybi] I-D Action: draft-ietf-hybi-websocket-multiplexing-00.txt

[Takeshi Yoshino <[e-mail]>]

Hi all,

Here's the initial version of multiplexing extension spec as HyBi WG item.

Change from tamplin-03
http://tools.ietf.org/html/draft-tamplin-hybi-google-mux-03 are:

- rephrasing
-- physical channel -> physical connection
-- logical channel -> logical channel and multiplexed connection
--- logical channel: logical data channels one for exchanging control
information and the others for transferring data for multiplexed connection
--- multiplexed connection: virtual WebSocket connection
-- multiplex control frame -> multiplex control block
-- uncompressed -> identity
-- some other issues pointed out by Greg's comments
- clarified that channel ID is assigned by client side on issuing
AddChannelRequest
- added EncapsulatedControlFrame block not to confuse intermediaries
- added diagrams for each multiplex opcode
- put description of multiplex control blocks into subsections instead of
using the list element
- replaced x-google-mux with mux
- sending a frame before waiting for AddChannelResponse (channel 1 is still
disallowed to do so) is now allowed for multiplexed channels
- added what intermediaries must do
- cite RFC 6455 in AddChannelRequest/Response description
- note about reuse of channel ID
- fail when AddChannelRequest is pointing already active channel ID
- AddChanneResponse with F=1: "Enc is ignored" -> "Enc MUST be identity"