Hi all,
Not really liking my previous missive regarding the Cloud
Profile too much (it goes too deep too quickly), please allow me to “pop
the stack” with a higher-level perspective, if I may. Forgive
oversimplifications and generalizations, but I am just trying to open up a
brainstorming dialog.
One might also note that since without SAF it is hard to do some
things it might be hard to find cloud use cases evolved and sophisticated
enough in the real world to be useful at this time in the industry. Just a
thought.
Existing and emerging management and security standards will do
things like the below (basically everything done for SOA, but more RESTful and with
a cloud provider and virtualization slant):
* Complain about misbehaving manageable resources (cloud
entities, perhaps, like virtual networks and CPUs) and provide metrics
* Provide protocols and schema to manage your cloud resources
(ask for a template, establish a contract, provision a VM, whatever). In short,
a sort of standardized equivalent to EC2 APIs and protocols, if you sould.
* Identity, autht, authz in the cloud (and federation stuff)
* Privacy and location stuff (ie: “don’t host my
data outside the country, etc.”)
* And so on…
So we shouldn’t do that stuff.
What is unique about SAF (some assertions)
* It’s not the Symptoms, it’s the catalog. What is
cool about the catalog?
A. TRANSLATION: You can “translate” a meaningful
pattern (Syndrome) to a meaningful action (Prescription).
B. CROSS-DOMAIN: This translation can be from one domain to another
domain or at a higher-level within the same domain. These domains can be IT/cloud
domains and/or business/application/service domains.
C. ITERATION: You can keep doing translations adding value over
and over again
D. EXPERTISE/KNOWLEDGE: The catalog can encapsulate the expertise
of human experts, which is often most valuable when it is doing (guess)
translation, is cross-domain, and is iterative.
What can we do with SAF (some examples that can’t be done
by any *one* of the existing/emerging cloud standards, I think:
* Integrate multiple IT domains
A. A security breach precipitates the deprovisioning of all
cloud instances that use a vulnerable template
* Integrate from application/business/level domain to IT/cloud
domain (kind of what I was striving for last Thursday). This is the catalog
expressed as business/IT expertise.
A. A business event from a SaaS company customer is translated
to meaningful actions making meaningful changes or requests to the SaaS
company’s cloud providers. For example, a 25% increase in a customer’s
sales may mean an anticipated 25 more instances provisioned by cloud provider
A.
* Integrate entirely at the application/business/service level
A. This is like our energy use cases, for example. An increase
in energy prices for one datacenter beyond a certain point might mean that other
datacenters, normally less preferred, are now the first choice for
provisioning. Ooops, this sounds like a cloud provider example. I guess that’s
because it is a *business/application* issue for the cloud provider.
Again, I’m just throwing this out in an effort to open a
brainstorming thread around value propositions of various scenarios.
Thanks,
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.