The Cloud Standards Are In
That settles it. We can put that puppy to rest. The entire problem is resolved.
It's no secret that I've been foggy about exactly what the Cloud is, who's behind it, and whether we are going to see major performance gains as a result of the potential for massive parallelism. And we've been a little, well very, skeptical about the Cloud hype.
Well, in Special Publication 800-146, the National Institute of Standards and Technology (NIST) of the US Department of Commerce has released their Cloud Computing Synopsis and Recommendations [PDF] in terms we all can understand.
This document is full of all kinds of nice goodies. There is an official definition of the Cloud, deployment models, constraints, services, scope, etc. No, we're not being facetious. We're believers in standards bodies and the NIST is high on our list. So their description becomes our de-facto standard.
Now the document does talk about grid computing, parallelizing components of the Cloud, and users' ability to take advantage of task parallelism. The document also alludes to applications having to be re-engineered to get the required performance gains. There is also a suggestion that what once was the prerogative of only the scientific community will now see the light of day in the new cloud architecture. But oddly enough, I still didn't see any promises of shortcuts in the process of parallel programming. There was no talk of paradigm shifting, or make-it-easier-innovation, nothing!
So with all of the advances and advantages of the Cloud, it appears that "somebody" is still going to have to do a lot of heavy lifting when it comes to parallel programming.
And we think it's safe to say that we all know who that somebody is ...