RE: Loops (RE: CPIM changes)

Adam Roach <[email protected]> Fri, 15 Nov 2002 15:28:58 -0600
Newsgroups gmane.ietf.impp
Message-ID <9BF66EBF6BEFD942915B4D4D45C051F3A64325@DYN-TX-EXCH-001.dynamicsoft.com>
> -----Original Message-----
> From: Dave Crocker [mailto:[email protected]]
> 
> Thursday, November 14, 2002, 12:34:26 PM, you wrote:
> Derek> If end-to-end security is a goal, then this is not true.
> 
> since I said "independent, heterogeneous" and since security 
> has such a
> problematic, variable implementation and use history, it is 
> pretty clear
> that security is not part of the gatewaying task.  It can't be.

I'm a little confused about the discussion of gatewaying between
heterogeneous protocols (which, if I'm not mistaken, is to meant
to denote "legacy" P&IM protocols). It seems to be completely out
of scope for the work that IMPP has been chartered for.

To my understanding, IMPP got where it is now through a series
of events along the lines of:

 - The IETF forms a WG to define Presence/IM protocols

 - The WG spends several meetings trying to agree on transport
   mechanism

 - Consensus never arrives, so the "council of nine" is formed to
   make a recomendation for a way forward

 - The council of nine, citing irreconcilable differences, proposes
   that three new working groups (PRIM, SIMPLE, APEX) be chartered,
   and each will define a different Presence/IM transport.

 - The IMPP working group remains to ensure full interoperability
   between the disparate protocols by having define core
   functionality and syntax for Presence and IM, as well as
   codifying base requirements to which the IM working groups
   will comply.

I'm sure someone will correct me if I'm straying too far off actual
history here -- but, assuming that the forgoing is a fair account
of how we got where we are today, the issue of gatewaying to
protocols that aren't under the aegis of the IETF doesn't really
seem to be in scope.

In particular, since IETF P&IM protocols will (or, at least, will
make a good faith effort to) comply with RFC 2779, the whole scope
of what IMPP is chartered to work on would seem to be constrained
to what has been defined (in this thread) as "homogenous" networks.

Right?

/a



  [reminder: [email protected] for non-technical discussions, please]