Cloud Computing Is Still In Its Adolescence

Compare the adoption rate for cloud services to that for virtualization, and it's evident we still have a ways to go.
The cloud is one of those amorphous technologies that gets trotted out as the answer to all of our woes, usually by people who don't think all that deeply about IT and its challenges. We hate to puncture anyone's bubble with a dose of reality, but at a macro level, adoption of all public cloud services except software as a service is going pretty darned slow.

For the past five years as part of our annual cloud survey, InformationWeek Reports has asked a simple question: What are your company's plans for cloud computing? The response we watch most closely is: We're receiving services today from a cloud provider. In 2008, 16% of survey respondents chose that option. In 2009, it was 21%, then 22% in 2010. It jumped to 31% last year, and to 33% this year.

Depending on your expectations, doubling in five years the percentage of IT organizations that use cloud services, and reaching one-third of organizations, might be pretty good. But compare the adoption rate for cloud services to another game-changing technology, virtualization. For almost every IT organization, virtualization isn't a matter of whether but how much. No one questions virtualization's core value proposition; the only question is about the breadth of applicability.

In contrast, two-thirds of IT organizations either have decided the cloud isn't for them or have yet to pull the trigger. The core value of the cloud is, in fact, in question.

It's not that IT planners are ignoring these services. The percentage of survey respondents who say they aren't interested has steadily decreased over the past five years. We asked the question somewhat differently in 2008 (we offered options for "don't know enough" and "no interest," but since then have offered only a "no plans" option). The percentage with no plans was about 50% in 2008 and dropped to 44%, 40%, 32%, and now 27% in subsequent years. That leaves about 40% in the planning and evaluation phase right now.

So while IT planners are interested in infrastructure as a service and platform as a service, they're having a tough time putting a value on such cloud services and coming up with an implementation plan that works for them operationally and financially. Here's the crux: IT pros see value in cloud computing--just like the not-so-deep-thinking pundits do. The problem is quantifying the value in a way that allows for apples-to-apples comparisons among providers (an issue we plan to tackle in the coming months), and IT pros view adoption as a significant change in the way they do business. These two concerns have kept adoption low.

Cloud vendors could help on both fronts. Pricing in particular isn't consistent; SLAs rarely help buyers sort out exactly what sort of service they'll get for the price they'll pay. Furthermore, cloud vendors haven't been much help in offering usable metrics and an interface to measure performance. That was actually one of the most positive points to come out of this year's survey: Whereas last year 37% of respondents said they didn't monitor performance at all, that percentage was down to 24% this year.

One reason we didn't see a bigger uptick in use is an increase in those who view cloud services as more risky than services from conventional providers. Last year, 44% (the largest percentage) said the risk was about the same; now 44% (the largest fraction this year too) say risk is higher with cloud providers. Both years, only 6% said risk is lower with cloud providers.

Slowing adoption is the requirement to create new IT processes to exploit the cloud (a risk in itself), along with the perception that these services themselves are risky. Literally hundreds of vendors are trying to create products that ease adoption--thus, in theory, lessening the requirement for new IT processes. But for IT planners, that just means more moving parts with more opportunity for failure.

The fact that hundreds of vendors are trying to solve these problems is in itself problematic. Before IT shops jump in universally, the way they did with virtualization, the cloud market will have to go through some growing pains, including vendor consolidation, development of better management and monitoring products, and a more sober assessment of the cost and value proposition.