[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

A Modest Proposal; was NIM BoF Minutes



Andrea, Walter, and other Friends of Comprehensive Unified
Modeling:

* A comprehensive model, be it NIM or something else, is
inevitable. It would be better if it happens now rather than
later. 

* Would a comprehensive model contribute to or reduce the
Tower of Babel syndrome? The process of developing it will
do nothing but help. Forcing everyone to use it before the
conceptual bugs are eliminated through simulations and field
tests, would cause confusion. So let's at least get started
on it now. I'm confident we can handle the deployment issues
when the time comes.

* But get started on what? what is the scope? Although it
would make life easier, not to mention more profitable, for
software designers and programmers to (re)use the same
objects (er, classes) for SS7/IN/TINA stacks as well as IP
stacks, the scope of a proposed IETF model can be limited
justifiably to the Internet. But that in itself is enormous
in comparison to what we have now, namely, the Tower of
Babel bandaged together by SNMP, SMI, MIBs and PIBs and god
knows what else. I don't mean to be negatively critical
about SNMP or its keepers. They've done a splendid job. But
the time has come for a more comprehensive perspective.

* Are we dealing with an information model or a data model?
Is the model a network model or something else? Is it
concerned with management or services or protocol
development? I suggest: All of the above and more. Toward
that end, I offer Thomas Magedanz's "On the integration of
IN and TMN -- modeling IN-based service control capabilities
as part of TMN-based service management", in _Integrated
Network Management_, Vol. IV (1995), pp. 386-397. My
recommendation is not that the IETF subjugate itself to ITU
modeling and implementation techniques, but that we
acknowledge the possibility that management and end-user
services -- and more -- can be modeled in a comprehensive
conceptual universe. Whether this actually happens is a
political/economic question.

* What might a global community of interacting Internet
objects (*) look like when they're up and running? Something
like this:


 Node 1   Node 2   Node 3 
 +----+   +----+   +----+     +----+   +----+   +----+ 
 | *  |   |    |   | * *| ... |  * |   |*   |   | *  |
 |  * |   | *  |   |  * |     | *  |   | ** |   |*  *|
 |*  *|   |   *|   |    |     | ** |   |   *|   |**  |
+-----------------------------------------------------+
|                                                     |
|                 Globally extended                   |
|             Virtual Internet Execution              |
|                    environment                      |
|                                                     |
+-----------------------------------------------------+
 |    |   |    |   |    |     |    |   |    |   |    |   
 | OS |   | OS |   | OS |     | OS |   | OS |   | OS |   
 | HW |   | HW |   | HW |     | HW |   | HW |   | HW |   
 |    |   |    |   |    |     |    |   |    |   |    |   
 +----+   +----+   +----+     +----+   +----+   +----+   
   ||       ||       ||         ||       ||       ||  
    -----------------------------------------------


The objects (*) are anything and everything that we're
talking about modeling: protocol objects on various layers,
management services, end-user services, applications, you
name it. Which ones you'd want to model in depth is an open
question.

Vie is composed of copies or kernels that run on the network
of Internet nodes such as hosts, endsystems, intermediate
systems, routers, etc. The OS/HW of each node is
unspecified, heterogeneous; the execution environment, Vie,
is the common element. It is available immediately through
Java/JVM/RMI without reinventing the wheel. Or one could
take the long way 'round like TINA and implement it through
CORBA. Either way, it can be done. (I suspect AS operators
will do the quick-and-dirty option through Java. But Vie is
virtual; it doesn't depend on implementation decisions.)

This is only scratching the surface ...

-- 
---------------------------------------------------
Tom Nelson Scott             Vedatel Co
1411 Sheffield Dr.           Bowling Green OH 43402
"In IP We Trust"   "Java Rules"   "E Pluribus Unix"
---------------------------------------------------



-------- Original Message --------
Subject: RE: NIM BOF Minutes
   Date: Mon, 21 Aug 2000 01:34:16 -0700
   From: "Andrea Westerinen" <andreaw@cisco.com>
     To: "Walter Weiss" <wweiss@ellacoya.com>, <nim@psg.com>

 Walter, I have a couple points on the notes.  The entire
discussion of scope - "What are we trying to accomplish with
a Network Information Model?" is relegated to a few comments
(mostly questions) in the "lengthy notes" section and the
statement that we need a "carefully defined scope".  I think
that this question was asked by several folks and was a
prime discussion topic.  How can a WG form without a defined
scope and charter?  Randy Bush stated this very well ...
"The reason this is a rathole is twofold.  The [modeling]
language [influences and] constrains how we think.  We
design the language to meet the needs of [what] we want to
express.  We need to discuss what we need to express, and
then design the models and languages."Also, the notes say
that general consensus was reached on the desirability "to
have convergence of the various information models in use
within the IETF".  I didn't hear consensus on this.  In
fact, many of the comments from the floor were quite the
contrary ("Would an effort like this contribute to a tower
of Babel or fix it?" and "I am not sure if this work will
help the various Working Groups; it may make them harder not
easier:  n --> n+1, n --> 1").    Regarding your 3 "locuses"
of positions, I strongly believe that there were 4 positions
- the last one being ... "Without a scope, I cannot decide
if the work has value or not."  This certainly was my
position and other comments below echo it.Lastly, a nit -
the session happened on Wednesday, August 2nd - not Tuesday,
August 2nd.Andrea

-----Original Message-----
From: owner-nim@ops.ietf.org
[mailto:owner-nim@ops.ietf.org]On Behalf Of Walter Weiss
Sent: Monday, August 14, 2000 10:45 AM
To: nim@psg.com
Subject: NIM BOF Minutes

Here are the NIM BOF meeting minutes. I would like to thank
the note takers (Steve Moulton, Glen Waters, and Bob Moore)
and Jeff for pulling it all together and providing a summary
for the faint of heart.

Regards,

-Walter

{snip]