RE: [rohc] Re: multiplexing SigComp and non-SigComp messages on t he same TCP connection

[zhigang.c.liu at nokia.com]

Hi,

Looking back in the mirror, the reasons why RFC 3320 became as it is 
today were:
- we assumed no security below sigcomp
- NAT prevented a flow-based architecture; so came the state model.

Now, looking at SigComp over TCP over secure channel (as we are
discussing now), I feel we are really dealing with a special 
sub-case than the general one we assumed above. So special that 
I'm wondering if it makes sense to define SigComp version 0x02 
for this case:

- it can utilize the security below sigcomp and the reliability 
provided by TCP.

- it can include support for multiplexing SigComp and non-SigComp 
messages (people seem to agree at least it is good to have).

- based on Version 0x01, with minimal changes.

Best of all, the new version number solves the negotiation problem
for "continuous mode" and multiplexing.

Any comments?

BR,
Zhigang

_______________________________________________
Rohc mailing list
Rohc@ietf.org
https://www1.ietf.org/mailman/listinfo/rohc