Virtualization Automation – A Proper Virtual World

Is virtualization entering a Dark Age?

December 16th, 2009

Cloud Computing is Pushing Virtualization into the Shadows

Clouding is the rage, everyone that is anyone is riding, or would it be better to say floating in, the clouds? I prefer to call it cloud bursting. Clouding is fine, but if it is a solution looking for a problem, then it is a horrible time and resources pit to fall into. No this article is not about the good or bad of clouding, but the impact of clouding on virtualization and the resources that design and support virtualization, and what is missing in virtualization to support clouding to a degree. Moreover of all the commercials about clouding that flood the airways now… IBM has one that just makes my teeth grind… IBM commercial says, if I heard it right… “…Cloud is Simple…” Well, nothing about clouding is simple. Pods are simple, from a hardware perspective; but the clouding software is not. Google has had what? Ten (10) or more years developing cloud oriented logic from a software perceptive!

Virtualization is now a commodity, just one more tool or component to a greater synergistic effort. But is it really? None of the complexity of the computing environment is reduced, it is increased. Just as virtualization made computing more complex, so does clouding, adding layers of complexity, dragging virtualization, provisioning, automation, reporting, and decision making together, of course decision making. Thus effective and efficient management and control are more important than ever. This is where things get dark, and everything fades to long grey shadows, and environment or datacenter architects run for the shadows, like roaches for cracks in the floor molding. Why? Because clouding is not easy, it is not uniform, it is not consistent. Cloud decisions are career changers… people resign, give up, or get nailed, by cloud solutions that do not live up to the hype, and no cloud solution, no matter how valued or vaulted lives up to all the hype. Thinking I am wrong? Dig a bit deeper, the evidence is there, that cloud design and implementation is hard work that some just cannot handle, or others took many years to get right at some effective level. Years? Oh man, that is a word that management hates with a passion, up there with corporate taxes.

Worse the cloud terminology is horrible. Whoever came up with the term Service? As in an application in a cloud is a Service? Talk about selecting the worse way to communicate to the end-user population a concept. Could it have been made any more, less informative? An application in a cloud is just that, an application! End-users understand applications, not generic terms like Service. Come on, call a stone a stone, call an application an application, for crying out loud. Better yet, call applications in a cloud… Solutions! End-users think in terms of problems and solutions.

The ugly aspect of clouding beyond the human impact is that the complexity of virtualization is not being acknowledged, and so talent and resources once dedicated to good virtualization solutions are bled down to the minimum, either let go, or feed to the clouding chewing machine. This is fact, not fiction. The push to have automation and autonomous systems mange a virtualization environment is a great goal, but is it a technical reality? Having reviewed more than 8 significant management, control, and reporting applications this year, including Surgient, Hyper9, ManageIQ, CapacityIQ, Liquidware Labs, etc. offers, as well as MOAB and Platform ISF. The word that comes to mind is… disappointment. Not because most of the these solutions in their own right have no merit, they do, but because all of them lack something that is critical to cloud busting, and some of us have been asking for as customer of virtualization for more than four (4) years, in my case closer to six (6) years, Predictive Analysis.

Predictive Analysis, in reference to virtualization, is the ability to do What If analysis scenarios against an existing environment. Look before leaping, in context to and virtualization models, which clouding must also address or deal with. For example, given ten (10) virtual machines, what happens when one (1) more is added to this situation? Saying, in simple terms, the result of such a proposed move, is good, bad or ugly? Now what is the impact if adding five (5) new virtual instances to an existing pool of a 1000? Predictive Analysis is applicable to every single customer of cloud computing, or bursting, large and small.

Never before, has predictive modeling been needed, especially in clouding. So where are the 100s of products that do predicative analysis? I would have thought that P2V would have pushed this gap into the light, from the dark shadows, but that did not happen, in part because P2V is not popular in some environments. So now cloud computing should do it, forcing predicative engines to factual reality. Talk about missing an obvious opportunity to establish competitive advantage! So, as virtualization is fades into the background, into a dark age, becoming more black-box than ever, and predicative analytical tools to supporting clouding own the light? Not yet, not yet.

Entry Filed under: A Proper Virtual World