David,
Excuse me if I am repeating myself but most, if not all, of the functions
you list below are well out of the scope of the current MSH specification.
That mailroom is a node with two MSHs and some amout of higher level
processing in between. To move on, we must achieve consensus that there
are no dumb intermediate nodes and no such thing as an intermediary with
only one MSH. Yes, code might be shared between the two MSHs but from the
view point of architecture, data structures, etc. there are two MSHs in
there.
Regards,
Marty
*************************************************************************************
Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown
Hts, NY 10598
914-784-7287; IBM tie line 863-7287
Notes address: Martin W Sachs/Watson/IBM
Internet address: mwsachs @ us.ibm.com
*************************************************************************************
"Burdett, David" <> on 09/13/2001 03:56:53 PM
To:
"'Dan Weinreb'" <>
cc:
Subject: RE: Use cases for IM's
Dan
You asked ...
>>>But then why does BM [the mailroom MSH] have to be an MSH at all?<<<
Some reasons include (there are probably more):
1. The BM MSH can keep the URLs of all external Partys which B does
business
with (e.g. A). This keeps all the look ups in one place which makes it
easier to maintain. Otherwise each application would have to do it on its
own.
2. It means that external businesses (e.g. Party A) can be given a single
URL to use to send messages for B, as the Mailroom MSH will forward it to
the correct application. This means that ...
3. If B re-organizes its systems internally and wants to move an
application
to a different URL, then it does not need to notify the external Parties it
does business with as the external URL does not change.
4. Some of the links to an application may not use an ebXML front end at
all, for example ...
A
--------------
BM
------
B1
-------
B
MSH
MSH
MSH
APP1
|
|
--------
B2
-------
B
|
MSH
APP2
|
-----
Inteface ---- B
App
APP3
In this case, B's mailroom MSH can allow the application (APP3) to exchange
messages with the rest of the world using ebXML which it could not do if it
had to use only SMTP. A variation of this is ...
5. The link to an application is done in batch, say once a day. However
there is a need to provide an immediater response over HTTP for messages
that are received. With this configuration, B's mailroom MSH can send an
acknowledgement or even a delivery receipt immediately back to the sender
(e.g. A) and later forward the message to the application
... then you said ...
>>>Since it doesn't actually interpret the message, there's no need for it
to know the ebXML MS protocol at all.<<<
Does it NEED to know it, perhaps not. However there is often going to be a
need for some kind of internal routing of a message from one MSH to
another.
You cannot just use a simple communications protocol as it menas the
parties
you do business with will have to be informed all of your updates. Using a
mailroom MSH makes it much easier to do.
... and then you said ...
>>>But if it [Commerce One] is still just passing the messages through
without interpreting them, the same point holds: don't consider it to be an
MSH.<<<
If you accept the benefits of using a mailroom MSH as described above, then
all we are doing is offering to outsource the service (for a fee ;).
David
