This is exactly the kind of explicit examination of pros/cons
that I believe we needed. Ultimately, based on the perspectives I’ve heard
shared, I think we have largely (but not entirely) paid the education/confusion
price. We are now left with an interesting and uniquely named TC, which is
good. Thus, I am leaning towards either keeping the current name or going the
complete abstraction route, as in “RAF - Reasoning & Automation Framework.”
I think the more interesting discussion, and really not so fully
addressed in this email thread, is the issue of adding a Cloud Profile to the
deliverables. On one hand, we gain specific relevance to a hot area, get to “test”
the abstraction of SAF against a specific domain that the current participants
are already involved in, and possibly get more synergies with other groups. On
the negative side, it adds to the work to be done starting perhaps with
carefully scoping the profile domain (do we do IaaS and/or PaaS, for example).
I assume that these will be put on the agenda either this week or
next week.
Regards,
Paul
Paul Lipton
VP, Industry Standards and Open Source
Member, CA Council for Technical Excellence
Phone (preferred number!): +1 215
539-2731
Mobile: +1 267 987-6887
Email: paul.lipton@ca.com
THIS MESSAGE MAY CONTAIN CONFIDENTIAL, PRIVILEGED OR OTHER LEGALLY
PROTECTED INFORMATION. IT IS INTENDED FOR THE ADDRESSEE ONLY. IF YOU ARE NOT
THE ADDRESSEE (OR SOMEONE THE ADDRESSEE AUTHORIZED TO RECEIVE THIS MESSAGE),
YOU ARE PROHIBITED FROM COPYING, DISTRIBUTING OR OTHERWISE USING IT. PLEASE
NOTIFY THE SENDER AND RETURN IT AT OUR COST. THANK YOU.
From: Marcelo Perazolo
[mailto:mperazol@us.ibm.com]
Sent: Monday, October 26, 2009 3:01 PM
To: saf@lists.oasis-open.org
Subject: Re: [saf] Possible TC Actions to Consider
I see the dilemma in either keeping the medical
analogy or going full-IT or something else... the name of the specification is
tied to this decision.
It depends a lot on what we want to be the focus. I believe we would want a
very broad core specification and then do a deep-dive in each domain
applications of the core spec, possibly in sub-groups under our TC or even
totally different specifications spawned in OASIS after the core spec is ready.
The IT Management space would be one, the Cloud space another, then others like
Process Management, specific industries like Energy, Healthcare, etc. All these
domains would have the option to extend the core spec to their own needs and to
define authoritative classifications types for symptoms/syndromes/prescriptions/etc.
(for better exchange of knowledge among intra-domain implementations).
A good non-biased replacement for "autonomic" has always been
"self-management" but then again "management"
has a bit more of an IT-specific connotation...
If we go with the IT-biased terminology my suggestion is this:
SMF - Self-Management Framework
AMF - Automated Management Framework
If we stick with Medical terminology:
SDF - Self-Diagnostic Framework
SDTF - Self-Diagnostic & Treatment Framework
If we abstract from everything and try to have a very broad terminology that
can be applied to any multiple domains:
SRF - Self-Reasoning Framework
RAF - Reasoning & Automation Framework
SRKF - Self-Reasoning & Knowledge Framework
In my particular opinion I think we should eliminate the IT-biased option. I
believe a medical analogy is something that everybody understands and despite
what blog comments elsewhere may say it's something that differentiates this
effort and gives it an interesting appeal... In a close second I also favor a
broader generic terminology with the words "Reasoning",
"Knowledge" and "Automation"... these are terms that - in
my opinion - fit well in several different application domains - but are not as
appealing I think.
I also disagree with having terms like Optimization and Remediation in
the name associated to the core spec because I believe the uses will be broader
than that - including prediction/avoidance and full root-cause resolution of
problems (not only remediation), security issues identification and prevention,
configuration and deployment issues, etc. In autonomic computing terminology we
used to classify "Self-Management" in four different quadrants:
Self-Healing, Self-Configuring, Self-Optimizing and Self-Protecting - symptoms
were supposed to help in all quadrants, not being specific to one or two of
them...
I'm also not too keen with "Framework" - I wish I had a better
suggestion for that, but at the moment I haven't...
Marcelo Perazolo
IBM Systems Director
IBM Corporation
phone: +1-919-5438491
email: mperazol@us.ibm.com
"Lipton, Paul C" ---10/26/2009 02:01:52 PM---Hi
all,
From:
|
"Lipton, Paul C"
<Paul.Lipton@ca.com>
|
To:
|
<saf@lists.oasis-open.org>
|
Date:
|
10/26/2009 02:01 PM
|
Subject:
|
[saf] Possible TC Actions to Consider
|
Hi all,
I took an action item to detail three possible
actions that should be explicitly considered, IMHO, rather than deciding “by
default.” If there is interest, perhaps this should be discussed at the next
meeting, make the decisions we need to, and then move on to the real work, as
they say.
Change the name of the base spec: Is the time for this past? I don’t know. On one hand,
both “Symptom” and “Autonomic” have caused (and are still causing) tremendous
confusion. If it is not clear why, I can elaborate, but I hope the reasons are
clear to us all. Heck, Symptom is not even necessarily the most important
entity in our information model! On the other hand, OASIS has promoted SAF a
bit and there has been some small notice of it amongst some respected people in
the standards community. It may be that we have already paid the confusion
price up front and now we can benefit from a fairly unique name. Symptoms
Autonomic Framework (SAF) is that, I grant. Either way, this TC should decide
if it wants to change the name. If so, sooner is better than later. A subtext
of this discussion is the commitment to a medical analogy, perhaps. Some
possible alternative names (better ones no doubt exist and should be
volunteered please):
· FOR: “Framework for
Optimization and Remediation” (simple and sweet or too simple?)
· FORAD: Framework for Optimization and Remediation Across Domains (in
the Urban Dictionary it also is a composite of “For Real” and “Radical”)
· FDT: “Framework for Diagnosis and Treatment” (possibly too medical
again)
The TC procedure, I believe, would be for us to
agree to all charter changes. Then we would have to make a motion for the TC
Admin to make a Charter Clarification Ballot. There is a 7-day ballot that
2/3’s of the TC must approve. Given the current size of the TC, this should not
be hard if we agree we need to do it. Once approved, OASIS updates the website.
We should add to the FAQ, IMHO, that the name change was for clarity.
Other possible changes:
· Remove medical analogy
and just go straight IT? Is this really
necessary or helpful? Not worth it if we don’t change the name of the TC.
· Add a Cloud Profile?
I admit that my initial thoughts are this is a good idea. Cloud is hot and SAF
could be uniquely useful to the cloud in complicated, cross-domain
applications, I think. It would broaden the scope of interest and at the same
time most of the folks involved are also interested in Cloud.
Of course, your mileage and perspective might vary
considerably.
Regards,
Paul
Paul Lipton
VP, Industry Standards and Open Source
Member, CA Council for Technical Excellence
Phone (preferred
number!): +1 215 539-2731
Mobile: +1 267 987-6887
Email: paul.lipton@ca.com
THIS MESSAGE MAY CONTAIN CONFIDENTIAL, PRIVILEGED OR OTHER LEGALLY
PROTECTED INFORMATION. IT IS INTENDED FOR THE ADDRESSEE ONLY. IF YOU ARE NOT
THE ADDRESSEE (OR SOMEONE THE ADDRESSEE AUTHORIZED TO RECEIVE THIS MESSAGE),
YOU ARE PROHIBITED FROM COPYING, DISTRIBUTING OR OTHERWISE USING IT. PLEASE
NOTIFY THE SENDER AND RETURN IT AT OUR COST. THANK YOU.