If you ask 10 technology experts to define cloud computing, you might get 10 different answers. If you don't know what it is, how can you expect to invest well? Well, here's your flash-card answer: Cloud computing is the computer as a utility.
While they lack monopoly power, the best cloud-computing providers, like utilities, leverage superior scale and per-use payment plans to offer resources at lower costs than what most companies could get on their own. Like utilities, no cloud service could exist without the necessary inputs -- in this case, hardware suppliers. At the other end of the chain, cloud users are restricted only by the limits of computers themselves, allowing for incredibly diverse opportunities.
I've put together a primer on the cloud that I hope will demystify an industry that often seems more buzzword than legitimate business model. Make no mistake -- it is legitimate, and there is great potential for those willing to dig deep. If you're curious about how to find and analyze cloud companies without needing a computer science degree, read on.
The computer utility defined
Utility computing was once offered only in long-term contracts with little flexibility, closer to the cable company than the electric company. Intel (Nasdaq: INTC ) found little success with this model, but Amazon.com's (Nasdaq: AMZN ) Elastic Compute Cloud, or EC2, later won the day by allowing users to pay only for what they used.
Many top cloud providers now offer per-use payments, though service-focused companies such as salesforce.com (NYSE: CRM ) may still use contracts. Contracts can work when pre-developed software is needed, ensuring ongoing support and regular updates. Pay-per-use usually is better for companies that have specialized needs and already have the people in place for high-level software work.
Intel is dominant in server processing and also competes in a crowded solid-state-drive market. I've repeatedly pointed out Intel's deep commitment to R&D, but that's only one aspect of the chipmaker's superior scale. Any search for the best cloud component investment should begin with Intel, but it doesn't have to end there. Other heavyweight hardware manufacturers also excel in this sector, as might be expected.
That doesn't mean that small companies can't succeed, but it's not easy. The entry barriers to profitably manufacturing computer hardware are quite high. It can be exciting to get in on the ground floor of a next-gen tech manufacturer, but without a deep understanding of what makes its technology better, such investments can be more like gambling.
Cloud-computing providers thrive off scale, which is not easy to reach. Very large data centers -- the only real source of cloud capacity -- may cost hundreds of millions of dollars to build, equip, and maintain. If a company doesn't have deep pockets, it probably won't get far as a cloud provider.
An average data center might have 1,000 servers, but massive operations usually have at least 50,000. Microsoft (Nasdaq: MSFT ) once spent half a billion dollars on one 300,000-server facility. A manager claims the operation costs Microsoft an eighth what most average data centers would spend per server over the long run. Scale allows these computing power plants to rent capacity at costs lower than those borne by smaller users while still netting a nice profit for themselves.
Sources: UC Berkeley RAD Lab, Microsoft Research, author's calculations.
The provider spectrum
But while most utilities are the same to consumers, cloud providers fall on a spectrum between infrastructure and platform -- not including storage, which is a separate but equally viable cloud model. Since storage is such a straightforward option, it's often part of the package when users need utility computing. Some cloud companies are storage-centric, but few are public.
Amazon's EC2 is all about infrastructure, and users can treat it like a big block of empty servers they might otherwise have in their own offices. Anything that can be done on physical servers can be done on Amazon's servers, with all the techie grunt work that entails.
Microsoft's Azure cloud, on the other hand, runs on the Windows operating system and provides pre-built applications, a midrange between pure infrastructure and pure platform. This restricts what programmers can do but provides greater support. Salesforce's Force.com is pure platform, offering limited modification options but excellent support.
Whether offering infrastructure, platform, something in between, or storage, it always pays to out-scale the competition. If a company commands enough resources, its profits are almost as guaranteed as those earned by utility companies -- provided users take advantage of those resources in significant numbers.
Software, services, and you
The final tier of cloud computing may be the most difficult to target for investment. It's certainly the broadest category. Most call it software as a service, just as infrastructure and platform options are dubbed services in the cloud. This tier is the only one most of us would ordinarily encounter.
The software that runs in the cloud ranges from Salesforce's myriad business apps to Netflix's online presence (now hosted by Amazon). Google (Nasdaq: GOOG ) is an app provider through and through, with a full array of cloud-hosted business solutions that compete with Microsoft's Office suite. The only real difference between software as a service and regular old software is the delivery method. Many investment considerations are the same in both cases.
Foolish final thoughts
There's a lot of ground to cover in cloud computing. I could never get to it all in one go. You can find out a lot more, though, in The Motley Fool's free report on the future of cloud computing. If you really want to discover what an incredible opportunity this is, our video report will give you a lot more information -- and some great tips -- on investing the cloud phenomenon. Thousands have discovered the cloud's potential through this revealing video, so join them now and get your copy of this free report.