Re: T2 - Assertions and Questions

[Dan Weinreb <dlw@exceloncorp.com>]

   Date: Mon, 13 Aug 2001 15:20:48 -0400
   From: David Smiley <dsmiley@mercator.com>

This is interesting.  So suppose we have end parties X and Y, and
they're communicating through three intermediarites A, B, and C:

X <-> A <-> B <-> C <-> Y

If I understand you correctly, one of these hops is considered the
"chasm"; let's say it's the hop between B and C.  Then:

   Assertion #2: 
   The To Party defines the location where messages intended for it are to be
   sent. 

In other words, when X and Y agree on a CPA, the CPA will assert that
the communication-protocol-level address for party Y is the address
that actually talks to C; and the address given for X is really B's
address.  Is that what you're saying?

   Assertion #3C:
   For our purposes, there is no such thing as a multi-hop message.

   Question #1:
   Can we eliminate any references to multi-hop or intermediate MSH from the
   spec?

As I understand it, you're not saying that messages actually never take
multiple hops, but rather than it *seems* as if they don't, "for our purposes".

So if I'm understanding correctly, the way I would phrase Assertion
#3C is "At the level of abstraction of the Message Services protocol,
there is no concept of a multi-hop message".  The interactions between
X, A, and B (and those between Y and C) are handled at a lower level
of abstraction, hidden from the higher level.  At the higher level, X,
A, and B together looks like a single entity; there is some lower
level that talks about how X, A, and B communicate among themselves.

Is that the idea?

-- Dan