Cloud computing: what it will take for start-ups to succeed

With Google's recent launch of its App Engine, and with the likes of IBM and Amazon having staked claims, cloud computing is clearly a major development in the IT landscape. The benefits are obvious, enabling enterprises to scale rapidly with a level of performance previously available to only the largest companies — all without adding equipment, software or staff.

It's early in the game — consistent definitions aren't even agreed upon in the industry. Is utility computing really in the cloud? What about managed services? The major companies in the category will certainly drive the definition of cloud computing. But as usual, expect start-up companies to disrupt it.

So far, the market entrants have played to their strengths. Amazon is leveraging its tremendous computing capabilities by providing customers with a virtual computing environment via EC2. Salesforce  is riding its software-as-a-service (SaaS) wave with Force.com — dubbing it "platform-as-a-service" — giving developers tools for creating business applications on-demand and without software. Pile on offerings from Sun, Yahoo, Ariba and more, all doing what they do best.

That still leaves plenty of room on the field for start-ups to do something different, which is what they do best. Unlike slow-moving market leaders, early-stage companies have the uncanny ability to identify beachheads adjacent to the market opportunities being established by the big boys. And more than that, they know how to move fast, growing right along with the niche market need they've chosen to address, turning that into a legitimate market which, in retrospect, looks obvious to everyone else. So what might be those beachheads in cloud computing?

To find their entry point, entrepreneurs will first need to determine what, in the enterprise infrastructure, can be cleaved off, or outsourced -- something the other players are not addressing yet. Cloud computing start-ups should work to find enterprise computing requirements that meet the following criteria:

* Computing needs that are not core to the business but are still required to support it. By that token, any enterprise computing requirement that relies on IT time, energy and budget for such processing needs as algorithms, heavy analytics and the like is one that enterprises would be likely to hand off to a vendor. For example, a financial institution calculating portfolio risk, or an enterprise trying to understand the optimal marketing mix for a campaign by using statistical and simulation analyses, could benefit from a cloud computing model.

* A cyclical business that has predictable spikes. Certain businesses that can predict their busy seasons with great accuracy would greatly benefit from a cloud computing model that allows them to scale up processing power only when they need it. An online retailer that sells gift items probably only exceeds capacity in the holiday season. An entertainment content Web site can predict demand for download bandwidth will increase with the video release of popular movies. These are the kinds of businesses that could benefit from overflow provisioning offered by a cloud computing start-up.