[dccp] spec changes as a result of packet processing algorithm

[Eddie Kohler <kohler@icir.org>]

Writing out the packet processing algorithm made it clear that we needed
some spec tweaks.

1. New state: PARTOPEN

   Tom Phelan mentioned this state a while ago -- The client has received a
   DCCP-Response packet, but has not yet heard anything further.  This is
   2/3 of the way through the 3-way handshake.  The client must retransmit
   the Init Cookie with every packet, and must include an Acknowledgement
   Number on every packet (no DCCP-Data packets allowed), until it hears
   again from the server.

2. New packet type: DCCP-SyncReply

   This packet is sent in response to DCCP-Sync packets.  It behaves just
   like a Sync, except that you don't send anything in reply.  In our
   previous DCCP-Sync-only document, it was hard to ensure that Sync
   exchanges terminated.

E

_______________________________________________
dccp IETF mailing list: dccp@ietf.org
list info:  https://www1.ietf.org/mailman/listinfo/dccp
wg charter: http://www.ietf.org/html.charters/dccp-charter.html