- "Cloud computers must operate on some sort of virtualization  
technology for many of the following features to even be feasible."

Really?   Why can't it just work on pre-installed capacity?    Blade  
arrays have been doing this for many years.

Certainly virtualization is important long term, but, in the open  
systems world at least, it's still _very_ new in the data centre.    
What happens if enterprises just aren't ready to adopt VMWare or Xen  
for a variety of technical or political reasons?    To paraphrase what  
one person here asked, "will you trust your hypervisor more than your  
OS?"

- "API for Creation, Deletion, Cloning of Instances"

I agree here, though I could see this as contentious, depending on  
whether people disagree on what an 'instance' is'.

BTW, wasn't this the point of the Open Virtual Machine Format (OVF)  
from VMWare, Xen, etc?  [ Of course, it's yet not ratified by the DMTF  
to my knowledge.   Oh, and also, DMTF... not exactly a well known  
standards body, yet. ]

- "Application Layer Interoperability"

It'll be interesting to see if user-packaging standards start to form  
here, like WAR files.    Or wouldn't a tarball be enough for many of  
these frameworks?

- "State Layer Interoperability"

There are billions of dollars thrown at interoperability, semantics,  
and integration annually by both vendors and customers, flowing like a  
river into a tarpit of epic proportions.    A large neon sign hangs  
over this tarpit that says "Abandon All Hope, Ye Who Enter Here."

There is some hope, of course - cue talk of the Semantic Web.  It's  
still too soon to tell what pieces remain to make it work, but a  
standard XMPP-based information bus is probably low on the list of  
priorities -- it's too specific a solution.  (The technical challenges  
in financial markets don't apply to every industry).   I'd say data  
model interoperability probably is a more important nut to crack.

- "Application Services".

Yup, I agree here.   But It'll be hard-fought street fighting to get  
the standards in place over the next few years.

- Automatic Scale (deploy and forget about it)

Funny, I keep looking for the "fast=true" parameter in my  
infrastructure, and I still haven't found it.

Joking aside, I do agree that 'adding & removing resources on demand'  
should be a feature of how one provisions cloud infrastructure. (Not  
coincidentally, that's one of Elastra's features in our preview  
release).   Though I'll go back to our conversation on performance  
management a couple of weeks ago:   many times, the best way to scale  
is to fix the application.

Barring that, the cloud would have to detect and rewrite the  
developer's shitty code for them.   This isn't completely out of the  
question, btw -- recent Oracle releases can rewrite a developers' SQL  
to perform better, and a DBA can choose to "swizzle" the statement at  
runtime without requiring a code change.   They added this feature so  
shops could workaround the crappy code in many packaged applications  
that somehow never seem to be fixed.

"-  Hardware Load Balancing"

This seems very much aimed at web application hosting.  While clearly  
important, that would be pretty limiting, especially if we're trying  
to shift the enterprise onto the cloud.  I'd go further and suggest  
that the network capacity and (virtual) topology itself needs to be a  
cloud service.

"8) Storage as a Service"

WebDAV?!   Why not AtomPub?

On another note, has anyone seriously considered iSCSI?  Certainly we  
can't think of storage just as web resources -- there also need to be  
mount points as a service, like Amazon EBS.    Or is this just a pipe  
dream?

Cheers
Stu Charlton
Elastra