Questions for the Cloud

Over the past year or so, I’ve been pretty breathless in my enthusiasm for cloud computing, and my enthusiastic writings around the topic have prompted a pile of reader mail containing many valid concerns over a possibly cloudy IT future.

I’ll admit that I’m an unabashed cloud cheerleader–I view it as an IT game-changer with the potential to dissolve enough of the friction associated with new technology initiatives to enable IT departments to act with more bottom line-enhancing agility.

You spin up just enough virtual infrastructure to serve a new project. If the project works out, fantastic. If it does not, you turn off the costs spigot by flipping the off switch on that virtual infrastructure.

If capital cost hurdles for new projects are lower, and if the sunk costs for failed projects are minimized, I think we’ll see more, better-tailored IT initiatives at companies of all sizes, which is good news both for organizations and for the IT industry.

Now, that’s the breathlessly enthusiastic vision, but if cloud computing is to realize that vision, then the horde of platform, virtualization, remote hosting, and software as a service vendors that are all hitching their wagons to the cloud are going to have to fill in the gaps that eWEEK’s readers have rightly pointed out.

Many of the reader concerns I’ve fielded boil down to wariness over surrendering control over organizations’ application and data to some Web-based business: How can I trust my business with these providers; what happens to my data if the provider goes belly-up; what happens when I wish to take my business elsewhere?

In order to allay the control concerns, vendors must demonstrate at least as much (and, to enable the fantastically agile IT scenario I sketched out, even more) data and application portability in the realm of the clouds as their customers now experience on the ground.

For instance, if an organization’s servers are cranking along on Amazon’s Elastic Compute Cloud service, and Amazon undergoes some system-wide outage, there must be a clear path through which those loads can shift to another piece of the cloud, be it some other EC2-like service, an on-premises data center, or some mixture of the two.

For software as a service providers, the portability question is even stickier, because a SaaS provider can not be parted from its software as simply as you can move a virtual machine from one host to the next.

Certainly, the the portability story for both of these infrastructure strategies remains a bit, um, cloudy to garner wholesale acceptance, and portability is far from the only cloud concern–security, offline accessibility, and regulatory requirements present additional challenges.

With that said, I don’t plan on handing in my Lando Calrissian Cloud City decoder ring any time soon, so I’ll be relying on our readers’ continued cloud-wary feedback to help keep me grounded.