Forrester Research analyst said at Cloud Connect you're guaranteed to save money in the cloud simply by scaling your workload based on demand.
(click image for larger view)
Slideshow: Top 10 Cloud Stories Of 2010
James Staten, the cloud analyst at Forrester Research, was the leadoff speaker list at Cloud Connect 2011, a UBM TechWeb event, Wednesday morning. His message at first seemed to be a defensive and negative one: "The power of the cloud comes from 'down' and 'off.'"
But for those familiar with the economics of cloud computing, you recognize in Staten's talk an amplification of Joe Weinman's 10 Rules of Cloudonomics. Joe was a cloud evangelist at AT&T when he composed the 10 rules; he now works for HP. But his work was original and a clear illumination of what cloud computing could accomplish back in an era when critics were sowing confusion about the cloud.
I've never studied the last nine rules of cloudonomics, but the first one has always stuck with me: You don't need to run your workload more cheaply in the cloud to save money. All you need to do is shut it off when you're not using it. No matter how cheaply you run an application in the data center, you will never get the cost down to zero, which is your charge in the cloud after you've shut it off. The way to get the benefit of the cloud is to use it when you need, shrink your use as demand decreases, and shut it off when demand has gone away.
"We design applications to be able to respond when the traffic goes up. It's also beneficial to you if the system shrinks when the traffic goes down," Staten said, pointing out an inverse advantage of "elastic" cloud computing. Many people turn to the cloud to get a guarantee of scalability when they need more horsepower. Why not also capitalize on its elastic property when traffic shrinks, Staten urged.
In fact, most application designers haven't thought about the consequences of traffic shrinking. They are trained to be prepared for success, capturing business from an onslaught of visitors and buyers appearing on the Web site or ecommerce system, and no one in the IT organization ever advocates turning an application off.
Thinking about traffic slowing down or going away is, well, defeatist, and nobody's paid well to think like that. Even though in the Internet era, what goes up must also frequently come down, unless you are Facebook.
What's needed, said Staten, is not an application whose different parts scale up by different means and at different rates. Rather, the user information requests, the buyers filling out forms, and ecommerce transaction processing must all scale at the appropriate rate, without being hindered by the other parts.
That means designing for rapid scale out and scale back, which means an app should be composed of core functions, each easily replicated when demand builds and shut down as it falls. "Down, off. Make it part of what you do in the cloud and you will win," he concluded.
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.