Software AG's Miko Matsumura provides a well-laid-out contrarian view of cloud computing in his most recent post. As Miko states: "If your cloud provider fails, then you fail. Unless of course your cloud provides nonessential services. Don't count on it. Large-scale failures such as Gmail recently point out some of the flaws in [the] 'don't worry your pretty little head' [promises]. You'd better start worrying your pretty head. Failures aren't the only problem implicit in the cloud, the lack of transparency can lead to privacy failure."
[ Get the no-nonsense explanations and advice you need to take real advantage of cloud computing in the InfoWorld editors' 21-page Cloud Computing Deep Dive PDF special report. | Stay up on the cloud with InfoWorld's Cloud Computing Report newsletter. ]
Miko is referring to the fact that "cloud," as a term and a symbol, has long referred to portions of the architecture you didn't want to explain. As Miko says, "Basically whenever someone didn't feel like drawing all of the network entities, they would just draw a puffy white cloud. In essence the puffy white cloud is shorthand for 'don't worry your pretty little head about this stuff.'" He's right about that, and also the fact that many considered "clouds" to be where the "miracle" occurs in the architecture. That was a running joke for years.
Miko provides a good representation of a small but more vocal minority out there: those that look at cloud computing as overhyped and too risky. They are correct to a degree.
Indeed, cloud computing is about leveraging IT assets you neither own nor host within your existing enterprise architecture, and of course you should carefully consider what architectural components should reside in the clouds. You know what? Failures are indeed going to occur from time to time. And you know what else? You must take a smart approach to security, privacy, and other issues, including performance and reliability, when considering cloud computing. We need to get over that.
Cloud computing providers are like any other type of technology providers: They provide opportunities and risks. You have to consider both carefully before making a transition to the cloud.
However, I do take exception to any public failures casting a negative light on cloud computing, such as the recent Gmail outage. Cloud computing systems will have outages from time to time, just as systems go down in your datacenter from time to time. You have to build a plan for these outages, just as you do with on-premise systems. Thus far, cloud computing providers have a pretty good uptime record; otherwise, you would have heard about it -- believe me.
I think that it's good to be skeptical about some of the promises being made by the rapidly growing, well-funded world of cloud computing. You need to be diligent around how you do architecture and including cloud computing within that architecture. There are no miracles here, but there are opportunities to improve existing IT around efficiency and cost. Your mileage will certainly vary, but you still need to get in and drive.