Avoid a Shakespearean cloud tragedy with production applications

Enterprises are investing in IT-operated cloud services -- so, what's the problem? Cloud services' value is moot without cloud-ready applications.

A cloud, a cloud, my kingdom for a cloud! Such is the focus of IT in 2014. But what use is a cloud if there are no cloud-ready applications at hand?

Infrastructure as a Service, or IaaS, and Platform as a Service, or PaaS, exist for one purpose: to host applications. And applications are where IT creates business value. While a well-designed and -operated cloud program does not create business value in and of itself, it can enable and support that value by hosting production applications in both public and private clouds.

Okay, so, what's the problem?

Consider this metaphor: King Richard (like the "application") died in Shakespeare's Richard III because he could not find a horse (or a "cloud") to carry him away from danger. In a reversal, there are many IT-operated cloud programs that have few or no production applications up and running. To go back to the metaphor, enterprises have horses, but no riders.

In short, 99% of enterprise applications were never built to run in a cloud. When you move your applications to a cloud, the process can be time-consuming and expensive, and it can potentially deliver little of the desired savings and operational efficiencies. Without the savings, the business often looks at cloud migration as a burden rather than an opportunity, which means that significant production apps aren't being migrated.

That is not to say that most of these applications won't run -- in fact, when you look at it by sheer volume, most will. However, the apps that generate the most business value are the ones most likely to underperform against expectations when they're moved to the cloud. This is because they are rarely built to take advantage of the new native advantages that clouds offer, such as elasticity, scaling and the economics of use-based cost models.

When you look at a typical enterprise application portfolio, you can see the 80/20 rule in play. In this case, 80% (or more) of the value created -- including revenue, competitive advantage and unique capabilities -- often comes from 20% (or fewer) of the applications. These 20% tend to be the larger, more complex and widely integrated applications. And most importantly, they have the most variable of demand curves, meaning high peaks and low troughs.

When looking at cloud economics (or "cloudonomics"), the applications with the most elastic demand curves are the ones to focus on first during cloud migration planning. The savings come from paying only for what you need when you need it, instead of all year long.

The imagined cloud value proposition

The typical enterprise deployment model is based on future capacity forecasting, plus a margin of error (sometimes as high as 50% or more), preprovisioned and left running 24 hours a day, 365 days a year. With business-critical apps, that can mean massive infrastructures with hundreds of servers; giant storage arrays; and redundant, high-capacity networks that often sit virtually idle for many hours in a day. Add in the infrastructure to support disaster recovery for these applications, and it's no wonder that average server utilization can be less than 10% in many enterprises.

Now, imagine an infrastructure that expands and contracts with demand, that can scale to handle the highest peaks of usage at 10 a.m., but is only a fraction of that size at midnight. When you pay for only what you use, the savings can be substantial. That's the core of the cloud value proposition for existing business-critical applications: paying for only what you use.

Most applications -- particularly the big, hairy ones -- were not built for the cloud.

The challenge is that business-critical applications aren't typically built for such elasticity. They were built to run on the preprovisioned peak capacity outlined above, and developers didn't need to think about scaling horizontally while maintaining or improving resiliency and availability.

So, instead of running your application on hundreds of internal servers operated and maintained by your IT operations teams, think of running it on the same hundreds of external cloud servers and paying for them 24 hours a day, 365 days a year. True, a compute hour on Amazon Web Services or Google, plus the related storage, may be cheaper in some cases than internally provisioned IT, but the cost advantage may not be significant when you add in complexity and management overhead.

And now, back to reality.

We have established two key and opposing points:

  1. Most applications -- particularly the big, hairy ones -- were not built for the cloud.
  2. It is the big, hairy, mission-critical applications that have the most potential value in cloud.

Do not despair, all is not lost. With a cloud modernization program that addresses and remediates the reasons these applications are not optimized, cloud deployment can deliver a significant return on investment.

About the author:
John Treadway is a senior vice president at Cloud Technology Partners and is based in Boston.

This was first published in April 2014

Dig deeper on Enterprise application integration with cloud providers

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchServerVirtualization

SearchVMware

SearchVirtualDesktop

SearchAWS

SearchDataCenter

SearchWindowsServer

SearchSOA

SearchCRM

Close