Defogging Cloud Computing: A Taxonomy

View Live Version of This Article

Michael Crandell is CEO and co-founder of RightScale Inc., a Santa Barbara, Calif.-based company that offers a cloud computing management platform, tools and services.

We’re heading quickly into the next big chapter of the Internet revolution, with tremendous buzz and excitement around the development of cloud computing. As with all major disruptive changes in technology, cloud computing has generated a flurry of definitions in the press and blogosphere, with acronyms and metaphors flying. At the same time, it’s important to remember that companies that are now deploying in the cloud have common problems they’re trying to solve, albeit in different ways using different approaches. We’ve found it helpful to create a mapping of these approaches –- a taxonomy of the cloud, if you will -– to make it simpler to understand product offerings and what benefits they provide for customers.

The term “cloud computing” has become a catch-all for any information technology solution that does not use in-house data center or traditional managed hosting resources. Self-defined cloud offerings range from Amazon Web Services and Google Apps and App Engine to Salesforce and even Apple’s new MobileMe service for iPhone 3G.

Among all these offerings is a common thread and key differentiator from the past -– the notion of providing easily accessible compute and storage resources on a pay-as-you-go, on-demand basis, from a virtually infinite infrastructure managed by someone else. As a customer, you don’t know where the resources are, and for the most part, you don’t care. What’s really important is the capability to access your application anywhere, move it freely and easily, and inexpensively add resources for instant scalability. When customers have the power to turn on and off 10, 100, 1,000 or even 10,000 servers as needed –- whether because a hot social Web application takes off, or a batch processing job starts to really crunch –- that is the core reason cloud computing is growing so fast. It represents a true democratization of Web computing, and it’s changing the way IT infrastructure is being delivered and consumed.

Our taxonomy of cloud computing –- which draws also on others’ work –- divides product offerings into three layers:

Today, we’re all witnessing the beginnings of a huge migration to cloud computing, simply because, for many applications, it is a better way to organize and manage IT infrastructure resources. Financial analysts like those from Merrill Lynch say there is a $160 billion addressable market opportunity for cloud computing. That’s a big number -– but not difficult to imagine. Even though we’re still in the early days, there are leading indicators that suggest how fast the cloud is forming. Forrester reported that bandwidth for Amazon’s EC2 and S3 in Q4 of 2007 exceeded all of the global Amazon.com web properties combined during their busiest time of the year. Take a look at Jeff Barr’s blog for an impressive graph of this data.

The technology industry has been moving toward open standards for some time, and cloud computing is the next logical step. Cloud solutions -– at any of the three levels described above -– are attractive for just about any company with an application that runs in a data center or with a hosted provider, that doesn’t want to reinvent the wheel or pay a premium. Multi-tenancy, low cost (metered hourly vs. monthly ), high availability with clustered servers (one goes down, spin one up automatically), virtually infinite scalability with a click –- all this is here, and here to stay. Our job as cloud vendors is to make it easily accessible and manageable, deliver best practices and continue to refine the architecture.

View Live Version of This Article