August 31, 2010, 12:32 PM — I often hear a debate about exactly who is using cloud computing - especially the IaaS variant. Many times I hear people aver that cloud computing is mostly being used by SMB companies. Others assert that cloud computing will really pick up when really large companies make the move. I dissent from both of those perspectives. I believe over the next couple of years it will be midsize companies that really leverage IaaS cloud computing.
Slideshow: Cloud Storage Lives Up to the HypeSlideshow: Six Cool Features in vSphere 4.1
Why would this be the case?
I think it's a pipe dream that small companies are going to really adopt cloud computing. The primary reason is that these companies are typically extremely short-handed in terms of technical talent. They've usually got a few overworked super sys admins fighting each day's fires with absolutely no time to invest in learning new skills.
And don't kid yourself, cloud computing requires skill upgrades. There's a world of difference between getting a single instance running in Amazon and implementing an elastic application surrounded by the necessary supporting services like monitoring and management. One small company sysadmin confided to me in a quiet aside - this was just a few years ago - that she was anxious about upgrading one of their core applications because the new version required an SQL database.
So I don't see small companies really adopting IaaS. Many people note that the low cost of cloud computing would be a good match for small businesses and assume that they'll be a ready market for the cloud. It's true that small businesses also run extremely lean and look to save every penny on IT they can. For this reason small businesses will be the home of SaaS adoption. SaaS is perfect for small businesses - they get the benefits of world-class infrastructure, enterprise-class features, and no capital investment. Frankly, I'd be surprised if the SMB market doesn't shift to a SaaS-dominated sector.
Looking at large companies, the barriers to cloud computing aren't a shortage of technical talent or a lack of capital. Far from it. Large companies are the ones that were able to build their own data centers and support a full complement of technical staff.
What holds back large companies is, in a sense, their success with the previous generation of computing. Because they could invest in the old model, they've now got an installed base of hardware and a large, top-notch technical staff on hand.
There's pressure on these businesses to justify the sunk cost of their hardware infrastructure, so they tend to more toward a vision of private cloud computing. And many of the talented technologists on staff see public cloud computing as a job threat, so there's a natural tendency to conclude that it requires further study and assessment and avoid rushing into anything too hastily. And in any case, many of these companies are executing five-year plans that have already been set in motion, so disrupting the current plan, even to consider something that could be very attractive, tends to be downplayed.
Midsized companies, on the other hand, could be the real beneficiaries of cloud computing. These are companies typically in the $250 million to $2 billion revenue range, and they commonly share these characteristics:
• They have technical personnel on board, but try and staff very efficiently. In other words, no fat. While their technologists are skilled, they're stretched, and there's never really enough of them. And if they have to hire fewer really good technology people, they want to hire ones that focus on business-oriented efforts (i.e., applications), not great talent that knows how to run infrastructure.
• They're capital constrained. It's no secret that the financial crisis has reduced the availability of loans, and that shortage has hit smaller companies disproportionately. And midsized companies want to focus their investment in areas that make a difference to what customers see, not in keeping the lights on.
• They may be growing rapidly. Or, if they're not in an expansion mode, they're trying to trim costs. In either case, a mode of computing that allows them to map resource use to business activities - while avoiding large upfront costs - is bound to be attractive.
