Re: call for participation

[Susan Hares <[e-mail]>]

Henry:

	I am a pragmatic, network operations person amd
IDRP code developer. I would recommend getting a 
mechanism going to pass changes that the IETF is
recommending to ISO protocols back ISO.  No politics,
no top dog in my mind - just the best technical
solution.

	Most of the changes to ES-IS for multi-cast come
from real life needs, and uses in the Internet.  There
are changes that I need to IS-IS to get pass third
party address so that we can make the next generation
NSF funded networks work with IS-IS.

	As to upper layers, I think that the thin stack
as it stands from Peter Furness's work is worth trying
out under applications.  Re-ordering of application layer
thought is going to come from DS-3, and OS-3 applications
and SCC.  Let's not stop anyone from doing good thinking.
We are only beginning to heal the thing of Connection vs
Connectionless.  The middle ground must be re-examined.


Do not think some of the IETF applications and protocols
could not do with a major re-write for IP Next Generation.
FTP has some real problems in moving to IP Next
Generation, and is an ancient FREE program.  

The current network technology is built on top a 
switch fabric that is LAN technology, bridges,
and large pipes.  Within a 12-18 months, I believe we
stand a large chance to see the ATM, and with
it circuit technology come back into major usage. 
With an underlying technology of circuits,
does the repeated Connectionless set-up make sense?
Let's think about that, not just say no or yes with
our past experiences and emotions.

I have an extremely pragmatic view of CLNP as an interim
step to a 15-20 year better solution.  I believe this
the ISO 5-10 year solution using CLNP with NSAP addresses
(20 bytes - watch for my Internet Draft on this!) is
pragmatic.  It will join user groups.  And with a little
squint - who says IPX can't be part of our family.
Why? - we need each of you and each person in the IETF to
think, dream, code, and try it out.  In-fighting, makes
this work not fun, and all of us think less.

Well, I'm not going to suggest any structure unless
you want to play in the IETF.  Then, I've chaired a
few committees there.  But I think a IETF/ISO blend is 
need to:

     1) Send IETF ideas back to
	ISO so we can get it to all possible users

     2)  If CLNP wins as IP Next Generation,
	  we'll need to have CLNP changes for sure.

	 I think it is the best pragmatic solution
	 since router's have it, lots of end systems
	  have it (Even Sun!)

     3) 10-15 year solution needs to create
	create solution out of connectionless/
	connection oriented to deal with 
	fabric changes

For those funding ISO activies, look at the words "public
domain".  A free program is more often used than one
that costs money.  Prototypes, "public domain" may have
had more impact on why IP has succeed where ISO protocols
failed.

(I'll gladly discuss this concept of free/payed off-line with
someone.  I will be glad to summarize
the discussion on-line after we've talked privately.
I'm got coding deadlines, and operational deadlines, and
I do not have time for more than this 1 long message.  I
have found private discussion between two parties leads down
really need paths of thought.  Groups tend to design
elephants.)

By the way, one of the things I like about IETF is pride of
ownership that RFC authorship has, and any pragmatism of
"just get it working" that still remains within in the
quieter majority. 


Susan Hares
[e-mail]
[phone] - for offline conversations too!

I work for Merit, Inc.  This is a personal observation
from my work and should not be taken to be the company line.