🚀 go-pugleaf

Both protocols are equally stupid when you consider the impact on the
client workstation.  And in the case of APPI, gee, IGRP was a great hit:
I can get it on  SO many routers.  (Yeah, right!)

Now the strategy that makes sense is to use OSF DCE or at least get IBM
to firm up their CPI-C over TCP strategy and then use them to build your
"blue" applications.

APPN should never leave the computer room.  Why should I support dual
stacks all over heck IF I get my blue connectivity over TCP?  And if I
DO decide I have to have APPN to support my IBM applications, then why on
earth should I use a tunneling technique rather than a native router? 
There is a reason they are called multi-protocol routers.  Tunneling 
causes a performance hit and is a kludge by definition.

Oh, and by the way, anyone still using IP only routers:  get in the real
world.  Sometimes business reasons force the issue.  You use the
protocol that makes sense in the homogeneous environment and TCP to bind
it all together.  I no more expect to remove SNA from my mainframes as I
would expect to remove DECnet from my VMS hosts.  They each provide
functionality that is important to those environment.  But that doesn't
mean I expect to implement DECnet, SNA, and TCP/IP on all my clients. 
Let the HOSTS take the burden of running the dual stacks.  Use TCP only
on the client as much as possible.

OK, enough spouting:  let the flames begin.

Charles.
--
Charles Ganzhorn			E/Mail:  ccganzhorn@mmm.com
3M IS&DP Network Services		AT&T:	 +1 612 736 7122
St. Paul, MN				FAX:	 +1 612 736 7689