Gartner released a report last week on 2008 IT spending that spotlights the next big thing: "per-use service-based models." This is a catch-all term that subsumes both cloud computing and SaaS. According to Gartner, "The projected shift to cloud computing, for example, will result in dramatic growth in IT products in some areas and in significant reductions in other areas." What Gartner elides is that cloud computing represents a fundamental shift in the role of IT: it removes from IT much of its daily work, but, and this is a big but, presents it with a different -- and more important -- set of tasks. Cloud computing outsources much of the dogsbody plumbing that has been IT's lot in life: provisioning servers, keeping the network up, providing 24\/7 infrastructure monitoring. Crucially, it gets IT out of the physical data center business -- no more special purpose space buildout, no more power calculations, A\/C maintenance. As Gartner notes, all this is outsourced to someone like Amazon. For those who are reluctant to trust their infrastructure to a bookseller, HP, IBM, and Sun are all active in this area as well. And let's face it, if you're a CIO for a medium-sized midwestern manufacturer, who's likely to be better at managing IT infrastructure, you or a technology company? In fact, this situation could be worse than I've just described: I recently talked with a financial services IT executive (financial services is generally considered the most technology-capable sector of the US economy), and he complained that his company couldn't hire A-level talent for traditional IT positions. Top talent wants to work for technology companies, according to him. So, on the one hand: crucial infrastructure that is challenging for IT shops to do well. On the other: technology experts ready to turn a labor- and capital-intensive activity into an outsourced, pay-by-the-hour service. To my mind, it's too logical not to come to pass. In this estimation, Gartner echoes Nick Carr, whose last book, The Big Switch, predicted the rise of utility computing. Carr, who has long viewed IT with a jaundiced eye, predicted that IT would abandon running their own IT infrastructures, much like early manufacturers quickly stopped running their own electrical generating plants when electricity providers came into being. Carr offers an example of the future by noting what a colleague of his did: snapped together online blogging software, flickr photo hosting, free wiki sites, and even more to create his site. All of this required no capital investment and no IT expertise. So hats off to Carr for limning the trend, presciently predicting it before Gartner made it a centerpiece of their analysis of the future of IT. I think he (and they) are onto something that will, thankfully, get rid of lots of thankless work for IT organizations. On the other hand, I think Carr is only half-validated by Gartner. He oversimplifies the world of IT when he compares it to snapping together online components to create an application. What his colleague did is fine for that person's purpose: to publish content. For example, I'm drafting this posting in Google docs and will publish it on the CIO website, which is powered by Drupal. No need for me to manage any computers or run a mini-datacenter. However, IT organizations exist to manage process, not content. What companies offer is not one-size-fits-all service, but customized offerings aimed at satisfying a customer want -- which is all about process. Even MacDonalds, the epitome of a standardized offering, moves heaven and earth to deliver one's burger, when one wants it, in a convenient location at a familiar-looking store. That's miles away from generic components. Consequently, while Carr is onto something, his prediction is superficial, and therein lies IT's future challenges. No longer responsible for drudge infrastructure, able to migrate work from talent-constrained operations groups, and freed up from pouring capital into undifferentiated IT infrastructure, IT faces fewer constraints, but greater opportunity, for the following reasons: Operations hasn't really disappeared, it's just different. While IT won't have to keep servers up and running, it will still be responsible for the company's applications. Given that Carr's vision of generic snap-togethers is too simplistic, IT will still have applications it is responsible for. So the challenge of IT administration will still be present. One could even argue it will be more difficult, since all the easy stuff will be given to Amazon to do. Now you really do need A talent. You can move the processing, but how do you move the data? Internal networks move at a minimum of 100 Mbs and often at 1000Mbs; on the horizon is 10Gps. While large enterprises can undoubtedly afford big bandwidth Internet connectivity, smaller organizations can't. There's a huge different between internal connectivity and external connectivity. Cioud computing will bring back the "last mile" problem with a vengeance. While one would like to be sanguine about the ability and willingness of telcos to see this as an opportunity, past experience does not necessarily raise one's hopes. This raises the bar for what IT needs to deliver. It's freed up from a lot of low-payoff chores, so it needs to deliver what remains even better. More crucial, however, is what accompanies to shift to cloud computing: the change in what constitutes the role of IT. Removing capital constraints means that IT should become more creative, more able to attune IT to business processes, more able to support innovation. Reduced costs implies lower barriers to use. Forget the received wisdom that cloud computing means more standardized apps run by a cheaper ops group. Cloud computing means cheaper innovation, quicker time to market, and more prolific experimentation. IT is more important in this environment, not less. Put another way, the constraint on IT moves from hard capital (server and data center space) to human capital (inventiveness, creativity, and innovation). Nick Carr got it half-right: IT is changing, big time. But he also got it half-wrong: IT is becoming more important, not less. Are you ready for your cloud computing future?