🚀 go-pugleaf

	Wow about a test of mail sourced routed to nixctc.de via nixctc-router
	via  unido.informatik.uni-dortmund.de.  Simple, right?  Perhaps you are
	expected to use an address like 
	
		pete%nixctc.de%nixctc-router @ unido.informatik.uni-dortmund.de

	Any one want to try it and see if my guess is right.


It seems the environment at unido is unnecessarly complicated.

	1. Mail from unido to uka is going via the uucp domain!  
           Clearly it would be preferable to us smtp/tcp/ip/x25.
	   If we could agree on a common network, instead of having
	   unnecessary hassels, we could run smtp with both unido
	   and uka.  

	2.  For the sake of "ISO-lation"  it seems that unido has
	    used a pseudo gateway "nixctc-router".  This is because
	    our x25/tcp/ip call manager was designed around the catanet
	    model (not a ISO-lation model), so it is assumed that the
	    network interface serves one net.  Since for the sake of 
	    "ISO-lation" we are kept on an "ISO-lated" net a pesudo
	    gateway was established (against my suggestion).  Now
	    as you can see our mail is being bounced around and returned
	    to senders.  This is expensive, I suppose the post likes that
            but I'm not very happy.  I hope we can get around all of the
	    "ISO-lation" stuff and just implement a simple catanet model.
	
	    As you can see, the unido smtp sendmail is clearly confused.
	    Someone forgot to tell him to rewite "nixctc.de" as "nixctc-router"
	    a then have "nixctc-router" perhaps re-write the address yet again.
	    How about following the KISS principle:

		Keep it Simple Stupid.

----------------------------------------------------------------------------------
       ----- Transcript of session follows -----
    While talking to unido.informatik.uni-dortmund.de:
    >>> RCPT To:<pete@relay.nixctc.de>
    <<< 554 <pete@relay.nixctc.de>... host or domain unknown
    554 relay.nixctc.de!pete... Service unavailable
------------------------------------------------------------------------------------------


	
	3. I don't think the charging algorithm for internet should be very difficult.
           The X25 call manager can keep accounting statistics IP packets in the  tables
	   with x121 and IP addresses in it.  A worst case algorithm, charging all packets
	   as going to the states for now would be acceptable.  Later, the charging algorithm
	  can be refined to do more accurate charging.  Of course, instead of using Nixdorf,
	  or Danet tcp/ip/x25 products which are software packages in the host and instead
	  using SISCO boxes this may not be possible.  I find the current "ISO-lation"
	  stratagy really stupid.


Pete Delaney - Nixdorf UCC	| pete@nixctc.de		Prefered Addr
Loffel Strasse 3		| pyramid!nixctc!pete		UUCP from Calf
7000 Stuttgart 70		| unido!nixctc!pete		UUCP from uunet
West Germany			| Phone: +49 (711) 7685-128