“Cloud computing” is becoming an increasingly popular term, but many people aren’t quite sure what it means. It refers to an important and long-term trend: computing over the Internet.
Hence it is useful to draw at least some rough but practical boundaries around the term and some of the computing trends associated with it.
What It Isn’t
Let’s first consider what cloud computing isn’t.
Cloud computing is sometimes applied to more or less any computing that doesn’t take place on a company’s own premises. This broadest definition thus encompasses traditional outsourcing.
This makes sense if you’re looking strictly at who is touching the computers, but it’s not a very useful window into how basic approaches to accessing and processing information are changing. Traditional outsourcing is largely about someone else’s taking advantage of economies of scale, best practices, and, perhaps, lower labor rates to do the same thing that you were doing before.
It’s an important trend, but it’s not a sea change in how we do computing–the promise that makes cloud computing so interesting. Outsourcing, therefore, is not cloud computing.
What It Is
That said, it’s useful to take a fairly broad view of cloud computing. It’s not just about the enterprise, the consumer, or delivering entire applications. Software as a Service (SaaS), Hardware as a Service (HaaS), Data as a Service (DaaS), and Web 2.0 are all part of the cloud. Even hosting providers are a sort of specialized, narrow case.
A broad view makes sense because all these intersecting sub-categories share at least one common characteristic: The network is the “abstraction layer,” the element that gives users a way into the Internet without burdening them with unnecessary technical details.
Software as a Service often gets used interchangeably with cloud computing, but it is really just a subset. Familiar business application examples include Salesforce.com (sales force automation) and Intuit’s QuickBooks Online Edition. These are essentially examples of familiar types of applications hosted by the application vendor and delivered as some sort of subscription instead of as a licensed program that end-users install on their own hardware.
The various online office suites from Google Docs to Zoho are more mass-market examples. Most familiar of all is perhaps the core application of the Web-connected world–search–especially, Google’s version.
Data as a Service is a subset of SaaS primarily focused on delivering data in response to a query rather than providing access to a more elaborate, full-blown application. Google search arguably falls into this bucket. StrikeIron’s Web services that verify and cleanse data are also examples, as are some of Amazon’s lesser known services that offer up historical pricing data and the like.
Hardware as a Service describes the much-discussed, if rarely seen in the wild, compute utility by which customers rent compute cycles as they need them. Amazon’s Elastic Compute Cloud (EC2) is one case; Sun Grid and IBM Deep Computing Capacity on Demand are additional examples that more explicitly target high-performance computing.
Vendors who offer compute utilities also tend to offer pre-packaged application loads for their target use cases. Thus, it’s misleading to think of HaaS as solely about hardware. As the category picks up momentum, expect to see many more bundled services and components complement raw cycles and bits.
Another important concept is Web 2.0 or the “read-write Web,” a term that emphasizes its contrast with traditional media that are far more about pushing information out than taking it back in. Web 2.0 includes services such as social networking sites, photo sites such as Flickr, and personal online calendars.
Although Web 2.0 largely began with consumer applications, both the applications and their ethos are increasingly seen in enterprises or at least touching enterprise applications.
As the above examples indicate, cloud computing is not about using networks and outsourcing to run the same old applications. It’s about new ways of accessing information, processing and analyzing data, and connecting people and resources across the network.
It is still in its early days–especially in the world of enterprise applications that evolves more slowly and deliberately than do consumer passions of the moment. But its implications for how we do computing and even who builds the systems and software are huge.
Gordon Haff ([email protected]) is principal information technology analyst at Illuminata.com in Nashua NH. An earlier version of this article appeared on the Illuminata.com Web site. Reprinted with permission.