ServiceMesh

In 1997, most companies had heard of the Internet and were running around trying to figure out how this suddenly-popular technology icon was going to impact their businesses. “We have to get on the Internet, NOW!” screamed many CIOs. While it was clear that the Internet was going to be big, Big, BIG, many companies were still struggling with an appropriate strategy for how to adopt the new technology and put it to best use. Every technology and business magazine was writing copious articles about the new Internet economy. Trade shows and conferences swelled with tote-bag carrying attendees trying to make sense of things. And there were more than enough self-proclaimed experts in the consulting and analyst ranks to help them.

Now, more than a decade later, cloud computing finds itself the center of a similar technologic whirlwind. Most companies have heard of cloud computing and sense that it could have a profound change on their businesses. There is plenty of reporting and analysis on cloud computing. But enterprises are still struggling to understand its limitations and how to put it to best practical use in their businesses.

In this article, we’ll attempt to clear the air and provide some grounding principles that will help you understand what cloud computing is all about, and how it will change your business in the future.

First, let’s deal with some terminology. Does “the cloud” really exist as a distinct physical entity? Larry Ellison, CEO of Oracle, in a famous Churchill Club interview, ranted that the cloud was a fiction, a mere relabeling of the entire existing network-connected global computing infrastructure. Describing Salesforce.com’s transition from being a SaaS company to a cloud company, Ellison quipped, “It’s the same reason that, well, Chanel last year was fuchsia and this year it’s become puce.” Continuing, Ellison said, “Our industry is so bizarre. They just change a term and they think they have invented technology.”

Is Ellison right? Is cloud just a fiction?

In one sense, he’s spot on. The cloud is not a revolutionary piece of hardware or software technology. When we’re using “the cloud,” we’re still using all the same networks, CPUs, RAM, and databases that we were previously. The real change is where those resources are located, who owns them, and how they are exposed for use by enterprise consumers. In other words, when we “move to the cloud,” we aren’t talking about a change of the fundamental hardware and software we’re using today so much as we’re describing a change in the economic and IT operating models surrounding the deployment and usage of computing resources. More simply, “the cloud” is mostly about changing our philosophy and behavior, not the hardware and software. Ellison hints at this during the interview but never explores it deeply.

Because we’re talking about a shift in philosophy and behavior, we believe it’s more proper to talk about the “the cloud operating model” than the “the cloud.” When one speaks about “the cloud,” it implies that there is a physical thing to be used, much like “the Internet,” rather than a new way of thinking to be adopted. Talking about “the cloud operating model” suggests a change in the overall IT operating model. In fact, we would even classify “the cloud operating model” as an implementation of a more general concept of an “Agile IT operating model.” An “Agile IT operating model” suggests a range of operating models that help enterprises deliver Agile IT, of which cloud computing is certainly an important operating model.

So, if cloud computing is really about the shift to a new operating model, rather than the purchase of a new “thing,” what are the characteristics of the cloud operating model? The cloud operating model is characterized by:

• Purchase of service rather than infrastructure – Instead of buying servers, installing them in a data center, loading them with operating systems and application software, and finally configuring them for use, we purchase computing services.
• Utility economics – Given that we’re purchasing services, we expect that we should play for the amount of the service we are consuming and we should stop paying for a service when our consumption stops. While the billing will have a lot of variability, ideally we’d like plans that have very fine granularity (e.g., minutes) with no long-term commitments. We simply pay for what we use.
• Self-service provisioning – If we’re going to achieve any amount of agility in our overall process, we need to eliminate all barriers between users and the resources that they wish to consume. The new cloud operating model should allow users to provision resources directly through self-service interfaces.
• Rapid, on-demand access to resources – Along with self-service provisioning, all resources should be available rapidly and on-demand. Resources should be live and available for use minutes after being requested.

The cloud operating model allows an enterprise to operate more responsively to changing business conditions. In return, this allows the enterprise to exploit rapidly changing, transient business opportunities.

One persistent question about “the cloud” is whether “private clouds” are a legitimate concept or not. Some have argued that “private clouds” are simply internal IT departments utilizing hypervisors and virtualization technology and are not really clouds. These people argue that only service providers can create clouds.

If we view the question from the perspective of the “cloud operating model,” however, things become more clear. It’s easy to see that the implementation technology is irrelevant. The important factor is whether the internal cloud is exposed to the rest of the enterprise as an on-demand service with self-service, rapid provisioning and metered utility economics. If enterprises run their internal clouds as captive service provider offerings, there is little room to argue that they are not clouds in the same way that external clouds. As the old saying goes, “If it walks like a duck and quacks like a duck, it’s a duck.”