Cloud -- in any of its forms -- is about changing the way you do IT.
No part of the IT organization appears to be able to fully escape the transformation to an IT service delivery model. Indeed, in this blog, I've talked at length about many of these changes: key roles and skills, organizational models, even the relationship between IT and the business.
But one important area has yet to be discussed: IT finance -- the way that organizations fund their IT investment.
And there are some interesting philosophical challenges shaping up here. Just to be clear, I don't have any easy answers this time :)
In A NutshellNo part of the IT organization appears to be able to fully escape the transformation to an IT service delivery model. Indeed, in this blog, I've talked at length about many of these changes: key roles and skills, organizational models, even the relationship between IT and the business.
But one important area has yet to be discussed: IT finance -- the way that organizations fund their IT investment.
And there are some interesting philosophical challenges shaping up here. Just to be clear, I don't have any easy answers this time :)
At an oversimplified level, this is the classic "irresistible force meets immovable object" scenario.
The "irresistible force" is increasing demand for variable IT services. Delivering IT as a service makes IT easier to consume; and, of course, anything that's easier to consume will cause much more of it to be consumed.
Indeed, there's plenty of anecdotal evidence that shows -- once friction is largely removed -- there's usually a bottomless demand from the business than involves more IT stuff.
This is not entirely a bad thing, if you think about it. The business is using these newer efficiently-delivered services to do all sorts of valuable and useful things -- like make money.
All is well, until you fully contemplate the "immovable object". Businesses, in general, like predictable costs. That includes things like headcount, R&D, and -- of course -- overall IT expenses.
As seasoned managers know, budgets are set far in advance, and it's not pleasant to try and adjust them upwards mid-course. More frequently, changing business conditions means that you end up getting your budget cut mid-flight. When the CFO is looking for quick cost savings, two groups tend to get routinely hit first: marketing and IT.
That's how the world mostly works. And it's not going to change just because we're all moving to a cloud model.
The Game Has Changed
In the familiar physical IT world, things were somewhat simpler to understand. You, as a business user, had "your" applications which of course ran on "your" physical infrastructure.
Most cost allocations, as a result, were pretty straightforward. The business was charged for the specific physical and software assets associated with "their" applications, as well as some shared allocation for facility, labor, etc.
Compare that with the new world of variable infrastructure consumption. Topics like transparency and chargeback appear relatively straightforward compared to considering the inevitable scenario when demand exceeds supply.
Who gets priority -- and who doesn't -- from the new shared "power plant"? Especially when there's not enough to go around?
Remember, in the old world, everyone had their own little puddle of dedicated-yet-inefficient infrastructure. If you needed more performance or capacity, you took out your checkbook and bought a bigger/faster bucket of infrastructure. If budget cuts were looming, "your" infrastructure was relatively safe, since the money was largely already spent.
In the new world, it's easy to see that annual IT infrastructure funding model might not be able to keep demand for variable infrastructure to support different aspects of the business.
And There's More
As this announcement from VMware reflects, we're starting to see new pricing models for software that assume a variable consumption model for software licenses -- so it's not just infrastructure stuff that's in play going forward.
And, if you read the subtext carefully, you'll see the word "elasticity" starting to be applied to more aspects of software. The emerging model here is different in a very important way that needs some serious consideration.
Today, most resource allocations for, say, an database task are rather static. Administrators assign fixed amounts of virtualized memory, CPU and perhaps I/O.
In the new world of "elasticity", a busy database task can request more infrastructure resources dynamically -- and, ostensibly, release them when no longer needed.
Forget, just for a moment, the potential challenges associated with simple pedestrian IT concerns (such capacity planning!) in this emerging world. Instead, imagine a rather important business application that decides it needs more resources right now -- and imagine that involves de-prioritizing other tasks on the same infrastructure right now.
Like perhaps your personal VDI session, for example :)
Welcome To The New World Of Resource Prioritization
The advantages of building IT to deliver shared services from a dynamic shared pool are impossible to ignore. We're not talking simple cost savings here; we're talking about agility and responsiveness -- something that every business leader craves.
The upshot will be that many IT leaders will be drawn into a relatively new and messy world of prioritizing business needs against available resources. In one sense, this is nothing new for the IT team.
What will be new is the frantic pace at which this dynamic prioritization is likely to demand.
Financial pressures will always incentivize an IT organization to run at ever-higher levels of efficiency against their shared infrastructure, which means that there will be less in reserve for the predictable "surprise".
But there are signs of hope ...
For one thing, the recent availability of compatible infrastructure from external service providers means that you can rent extra capacity -- quickly -- if needed. That's a helpful development, if you think about it.
And, of course, as technologists we can imagine new variations of management tools that will help us understand and predict the various demands against the new shared infrastructure (think analytics), create categories and rules for who gets what resources under different set of circumstances (think policy management) and report back to the business the actual service levels delivered.
One thing won't likely change, though.
Getting the business to pay for the IT services they need and want -- well, that's still going to be a challenge for the foreseeable future :)
By: Chuck Hollis
