RE: m3ua status

My 2 cents about this and related issues:

What I think from technical point of view:
1) IMO, SE-IPSP model is a true peer-to-peer model. I don't see any
architectural deficiencies but OTOH I think existing documentation is
unsatisfactory.

2) I don't think DE-IPSP model is a good approach.

3) I think SG-SG model is useful, because it provides relaying capability.
It can be used for IPSP-like communication between IP-resident peers as
well.


Now, what can be done:

a) A draft clarifying IPSP usage as defined by the current document. This
should be a self-sufficient document, should not change something
architecturally but should include all details which are discussed in this
list throughout the years -the current document unfortunately doesn't-.

b) Revisiting SG-SG draft. I will submit a new version soon -in 5-10 days-
still as a non-WG draft. There has been some off-list comments about it
during the last 2 years or so, but architecturally it won't be any different
than the version Brian posted the links to.

Where I really am not sure what to do is, do we want to provide multiple
choices for IPSP-model? Can we have significant changes and if yes to which
extent? We already have SE-IPSP and DE-IPSP, and as indicated above, SG-SG
could serve the same purpose as well. More choices may be confusing for
people. My personal preference would be -considering both technical and
non-technical aspects of the problem-:

i-  Keep SE-IPSP but not DE-IPSP.
ii- Define SG-SG, by emphasizing its use for relaying rather than IPSP-kind
of operation.


    Thanks,
    Tolga
  -----Original Message-----
  From: [email protected] [mailto:[email protected]]On Behalf
Of Ong, Lyndon
  Sent: Friday, November 11, 2005 5:41 PM
  To: [email protected]
  Subject: [Sigtran] m3ua status


  Hi Folks,

  I have been watching the mounting pile of emails on IPSP procedures with
some concern.
  I think at this point it is quite difficult to make any technical changes
to the M3UA bis
  spec, as this is well past the review period.

  At the same time, it seems clear that there are major differences in
peoples' interpretations
  of some sections in the text.  Continued discussion on the email list does
not seem to be leading to a
  consensus, either.  Do we need an actual meeting, or some more organized
effort to
  reach a conclusion?  I'm open to suggestions here (can be sent off line if
you wish).

  Thanks,

  L. Ong

_______________________________________________
Sigtran mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/sigtran