It’s tough to get a handle on cloud computing these days. Everything that has even a vague correlation with either computers or networks is a "cloud" in the eyes of eager PR types. Despite the hype, it's fair to say that cloud computing is the most potentially transformative issue in both networking and computing, and that would mean that if it fulfills its promise, cloud computing could become one of the top three transformational issues of business technology. It's almost certain that every single business that uses computing will use some form of cloud computing in the future; the big questions are "How much?", "What for?" and "How do I get there?"
You'd think that the cloud was the future, and it sort of is as long as you’re careful to define what that means. In our spring business survey, we found that on the average, businesses expected that at most about 23% of their IT spending would migrate to public cloud services. Nearly all large businesses (the top 8,500 companies worldwide) believe they will employ some private cloud technology. The lower tiers of companies are unlikely to do anything with private clouds, unlikely to commit more than a third of their IT spending to the public cloud, and are therefore to an extent staying where they are now.
To an extent--but recall that nearly every business will do some cloud-sourcing. For the business of the future, the cloud is a reality, but at the same time likely a minority in terms of IT commitment. That makes getting there and managing the end-game all the more problematic. The special concern is whether there's any way to predict how cloud commitments might start. From that, their progression could be clearer and the pace of likely evolution easier to predict.
Enterprises have pretty well-developed views on cloud-source application priorities. Personal productivity applications--what could be called "Microsoft-Office-as-a-Service"--is the area where most businesses are interested in the cloud. A close second is communications and collaboration applications including email, UC/UCC, etc, and the third place is departmental administration applications like sales force management. About two-thirds of all businesses are likely to use the cloud form of at least one of these by the end of 2011. In contrast, only about 18% are likely to have done any kind of cloud-sourcing of their core business applications in that period.
It's probably not a coincidence that these early cloud applications are ones that neither draw from nor contribute to the traditional corporate databases, and that they don’t require integration with core business applications. In my spring survey, in fact, fully half the enterprises planning or using these productivity applications said that their "contained" nature was one of the prime reasons for starting there. But this immunity from integration issues may be illusory in the longer term, particularly for collaboration-as-a-service offerings.
The great majority of business collaboration today takes place around data elements drawn from core business applications. Getting help with an account or order, responding to a customer inquiry, and reviewing an unusual inventory trend all demand that the people involved get on the same page--literally, in a display sense. This means that at minimum any cloud-based collaborative tools have to be able to synchronize, in some way, the application views of those involved. It might be a matter of mashing up displays, or it might require deeper integration through application program interfaces (APIs), but in either case the cloud application and the company's own IT tools now look unified. How unified they really can be made to be should therefore become a consideration in picking a cloud productivity application.
Today, that's not likely to be true. Enterprises are often climbing a thin vine to the cloud, not because they want to, but because they haven’t yet figured out what their overall cloud strategy will be. Absent a policy for the cloud of the future, they can't tell if the cloud choices of the present align. This is particularly likely to be true where cloud projects go forward without IT supervision or even involvement. A cloud project done independent of IT is five times as likely to fail as one run by IT, and three times as likely to fail as one where IT advises.
This isn't an unusual story, either, since failures in basic cloud policy planning are listed as the number one cause of cloud project failures. The moral is that it's critical for enterprises to formulate a strategy for cloud computing that recognizes the inevitability of hybridizing public cloud applications and resources with private IT, whether that private IT is based on private cloud technology or on the current data center client/server model. A plan that starts with hybrid cloud requirements can build cloud services into that plan and avoid creating integration, management, and security problems down the road.
