ctw@aerospace.aero.org (Charles T. Wolverton) (11/03/90)
We would like to use an IP internetwork as an intermediary between
two OSI networks for the purpose of X.400 E-mail exchange. The
following kludge uses Wollongong products and appears
(conceptually) to work:
------- -------
| MHS | | MHS |
| --- | | --- |
------- | ULS | | ULS | -------
|X400 | --------- |w/RFC| |w/RFC| --------- |X400 |
|MTAs | | TSB | |1006 | |1006 | | TSB | |MTAs |
| ON | |-------| |-----| |-----| |-------| | ON |
| TP4 | |LLS|TCP| | TCP | | TCP | |TCP|LLS| | TP4 |
------- --------- ------- ------- --------- -------
+ + + + + + + +
+ + + + + + + +
++++++++++++ ++++++++++ ++++++++++ ++++++++++++
OSI #1 TCP/IP #1+ +TCP/IP #2 OSI #2
+ +
++++++++++++++
TCP/IP internet
Each OSI/TCP #N logical pair is physically one network. The
Wollongong components are:
TCP-based MTA: WIN/MHS, WIN/ULS (using the transport switch in RFC1006
"position"), WIN/TCP
Transport bridge: WIN/TSB, WIN/LLS
Altho shown logically separated, hopefully each MHS/ULS/LLS/TCP set would run
on one machine.
The assumptions leading to the kludge:
i. The WIN/ULS transport switch is static, not dynamic -> a simple
dual stack ULS-over-LLS+TCP wont work -> need the TSB
ii. The TSB needs a true destination host address, not another TSB
-> need the two TCP-based MTAs (MHS/ULS/TCP)
Questions:
i. Is there a more straightforward way to do this??? (with
commercial products - no ISODE, "you can easily hack the blah, blah
code", etc. - we're buyers, not developers)
ii. If not, are the above assumptions correct?? Should the kludge
work?? Has anyone tried it (or something similar and, hopefully,
simpler)??
Many thanks for any help. Also, apologies if this isn't the best
list for the question.
-chas