Re: CTP and CARD Drafts

Hi Allison,

> So, I've done a little talking and looking at the document.  The example
of a
> service that was supposed to be covered in Appendix C is missing.  There
is no
> companion document describing a use case.  What's up?
>

At the Vienna meeting, there was a lot of discussion about the need for
context transfer for header compression. The ROHC experts seem to be divided
on this, and predictably the opponents of context transfer from ROHC came
out in force in Vienna to shoot down the proposal to include header
compression as an example for CTP when they heard it was on the agenda.
Since it really isn't a Seamoby issue whether ROHC agrees on the need for
header compression or not, it did not seem fair nor worthwhile to hold up
the CTP draft until ROHC came to concensus about the need for header
compression. The WG, noting the desire of the IESG to quickly finish the
work and close down Seamoby as soon as possible, decided that Rajeev Koodli
and Carsten would pursue a ROHC context specification in parallel, and the
CTP draft would include a simpler example, on which everybody could agree
that context transfer would make a difference. That example is in Appendix
B, multicast listener context, and it contains some calculations which
illustrate how CTP could improve handover performance. Note that this
doesn't mean the people who thought there is value in header compression
context transfer have now given up that view. Rajeev has done some excellent
experiments that indicate a significant benefit from header compression
context transfer, and he was planning on presenting the data at the MOBOPTS
meeting in Seoul, but we ran out of time. I think the feeling now seems to
be that MOBOPTS would be a better context (so to speak) for discussing the
header compression context transfer, since determining the value of context
transfer for header compression requires carefully designed experiments
rather than loud microphone action.

So, I think some in the WG may be concerned that approval of Seamoby CTP for
 publication could be held up pending a resolution of the dispute within the
ROHC WG about the value of context transfer for header compression. The
multicast listener context transfer example in Appendix B was supposed to
provide the use case you are looking for. Do you think that is sufficient or
should there be more?

            jak