Print

Print


Hi Leonid,

My interpretation of Robin's response to you is that she is thinking VSM could be used internally to the Oracle product line to detect emerging performance problems in the database management system itself (hence the reference to database administrators (DBAs) and performance tuning).  This would represent a product-level application of VSM which is, I think, very different your initial intent.  

My understanding of your initial intent was there be some user-configurable set of enterprise-specific controls (homeostats) reflecting the critical elements of VSM (referred to as cyberfilters elsewhere) and that they would sit on top of Oracle (or SAP, or whatever) as a software-application level implementation (possibly as a middleware layer?).  Is this accurate?

If it is, you will need to reset the terms of the conversation with the Oracle user group rep (Debra).

Best regards,

Stephen Davies

-----Original Message-----
From: Forum dedicated to the work of Stafford Beer [mailto:[log in to unmask]] On Behalf Of Leonid Ototsky
Sent: Saturday, October 04, 2008 4:46 PM
To: [log in to unmask]
Subject: [Fwd: Stafford Beer] and Oracle

Colleagues,
It is a reply from Oracle UK

Leonid - http://ototsky.mgn.ru/it
------------------ Forwarded -------------------
Subject:   Stafford Beer
From:       "Lilley Debra" <[log in to unmask]>
Date:   , 05  2008, 00:42
To:   [log in to unmask]
Copy: [log in to unmask]
            [log in to unmask]
--------------------------------------------------------------------------

Leonid,
I know I was not able to give you the answer you were looking
for before but I have not forgotten you. I have been talking to friends
in the Oracle World and was introduced to Robyn who was good enough to
look at what you sent, and this is her response:



My 'message' is that variance normally exists in any system.  Some
variance is controllable, some is not but by designing our systems with
the expectation of variance, we can make them more resistant to
variation and therefore more robust.  My ideas on this come primarily
from Genichi Taguchi, but there's a smattering of other concepts from
classes in operations management/research and quality engineering in
there too. But I'm really just advocating that dba's approach the
management of database systems systematically and that they measure the
variation in system performance, as I think users are happier when they
know what to expect.

The level of change that Leonid seeks would be a much longer term
project and I think it would require involvement from Oracle. I bristled
a bit at section on alerts when the system is failing to achieve
performance - if anything, I think our industry relies on sending alerts
for problems rather addressing root cause and eliminating them.
However, can software be built to receive it's own alerts and adjust
itself accordingly?  If I understand correctly, this is the ultimate
goal for VSM which is essentially an artificial intelligence system.  (I
studied this a bit in school too - I had to build a small AI system for
a class, I'd forgotten all about that experience.)  While AI has made a
lot of progress since then, a system can only know what we tell it so
from my perspective, and Dr. Taguchi's, there will always some bit of
unknown and uncontrollable variation.

I would wager that somewhere inside of Oracle someone is already looking
at VSM or something like it. I've been following Graham Wood's
presentations for a while now as it seemed that with ASH and DBTime,
Oracle was beginning to provide the means to measure variance.  As for
Leonoid's quest, Oracle is advertising a self-tuning system, which could
be considered a step toward Viable Software. It would be interesting to
find out if any of Oracle's work had been inspired by Stafford Beer or
Dr. Herring's paper.  A methodology SIG, as you mentioned in your
response, could be helpful but I think it would need participation from
within Oracle to make it work, as Leonoid's area of research would need
to be taken into consideration at the earliest stages of design.

All in all, it's an interesting set of ideas, there is probably
potential in them but it's a big change and one that would need to come
from within Oracle.  Plus, there are many operations research theories
out there, a project like this should begin by identifying problems,
evaluating multiple methods and then determining which approach would
have the greatest likelihood of resolving them.



I haven't found a methodology SIG but still looking. Robyn and I are
meeting later this month and the Oracle gentleman she mentions will also
be there so perhaps we will be able to find out what Oracle is thinking.



I have seen your post on the web about my original response and am happy
for that to stay but would appreciate if you could remove my contact
details I am receiving an awful lot of Russian spam which may not be
connected but I would feel happier.



Fusion Applications are now well into development so perhaps things have
moved on from your original request but Fusion Applications is as much a
concept for a single user experience so any component in the future will
have a design element.



Debra Lilley

Deputy Chair - UK Oracle User Group

25 Years of Serving the Oracle Community

IOUC Product Development Committee & Fusion Channel



Mobile              07867 822186

E-mail           [log in to unmask]
<mailto:[log in to unmask]>



------------------------------------------------------------------------
----------------------------------

For information on UKOUG visit www.ukoug.org <http://www.ukoug.org/>

------------------------------------------------------------------------
------------------------------------



Although this email was sent from my employer's mail system it should be
seen as coming from UKOUG.






~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For more information go to: www.metaphorum.org

For the Metaphorum Collaborative Working Environment (MCWE) go to:  www.platformforchange.org

METAPHORUM eList Archive available at - https://listserv.heanet.ie/ucd-staffordbeer.html

Archive of CYBCOM eList available at - http://hermes.circ.gwu.edu/archives/cybcom.html
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For more information go to: www.metaphorum.org

For the Metaphorum Collaborative Working Environment (MCWE) go to:  www.platformforchange.org

METAPHORUM eList Archive available at - https://listserv.heanet.ie/ucd-staffordbeer.html

Archive of CYBCOM eList available at - http://hermes.circ.gwu.edu/archives/cybcom.html
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~