Gorry, This version adds the following in section 3.8: The purpose of this IANA-assigned port is for the operating system or a framework to receive and process DCCP-UDP datagrams for delivery to the DCCP module. Because of this, the IANA-assigned port SHOULD NOT be used as the Destination UDP Port by a DCCP-UDP server listening for incoming DCCP-UDP packets and SHOULD NOT be used as a Source UDP Port by a client application sending DCCP-UDP packets. I think this is saying that a system-wide DCCP-on-UDP encapsulation service SHOULD use the IANA registered port, while other applications that implement the encapsulation SHOULD use ephemeral port. The phrasing in the draft is very confusing though, and can be interpreted as meaning that the IANA-registered port can't be used at all. Colin On 10 Mar 2011, at 21:21, Gorry Fairhurst wrote: > I've posted a version of the draft that addresses all the comments I know how to address. It includes the following new text: > > * Small NiTs as noted in WG list. > > * A note in the intro saying you can use this with a UDP-Only stack. > > * A small intro to 3.3.1 to clarify partial checksum and NAPT. > > * A note on issues when minimal length ICMP messages do not tell you the full port information. > > * A note flagging the DCCP-UDP reduces the MPS. > > * A note clarifying (I hope) the use of well-known ports. > > * Section 5 says when you can do both DCCP-STD and DCCP-UDP you SHOULD choose DCCP-STD. > > * Example for SDP added. > > > Two issues are not presently addressed: > > * How to do ICE - intentionally left as other work. > > * How to say DCCP-UDP if this is the *only* DCCP stack that is supported, although the specification says currently that the presence of "a=dccp-in-udp" conveys no information about whether or not the offerer is listening for DCCP-STD connections, it does not go any further. > > We plan to ask people in the MMUSIC WG if the SDP description will suffice, or whether there is a simple alternative. Please also respond to the DCCP list if you wish to discuss this. > > Best wishes, > > Gorry -- Colin Perkins http://csperkins.org/
