Application scaling is easy, except when it isn't

How to make dynamic scaling of an app actually work

When properly deployed, an application running in the cloud is going to be available -- regardless of how many users need to access it and whether a failure somewhere in the system takes down one copy.

It's that scalability and resiliency that draw IT decision-makers to cloud computing. But dynamic scaling isn't a simple thing, or at least not as simple as admins would like it to be.

An application needs to be suitable for scaling, and one app's requirements won't be the same as another's. And for autoscaling to be effective and efficient, admins will need to plan carefully and determine what amount of CPU use is acceptable before adding another server to support an app.

Veteran IT architect and engineer Mike Pfeiffer breaks down these and other key determinants in this handbook's first article. If you can achieve dynamic scaling of an app -- either manually or automatically -- that will go a long way to keep that app available and cost-effective. But, Pfeiffer cautions, these outcomes aren't guaranteed.

You'll need the right app governed by the right rules. Admins should consider carefully which apps belong in the cloud and how they want them to behave when they're there.

A business can certainly reduce its application infrastructure costs by scaling in the cloud, but only if it smartly plans and manages the necessary resources. The beauty of the cloud, after all, is that you pay only for the services you consume. This makes it all the more important to be prudent about your decisions with application scaling -- especially with a flexible app -- so that you match demand rather than guessing at it.

Dynamic scaling is easy to do, but that doesn't mean it is easy to do right.