Exerting Control Over Your Sprawling Cloud

We all know that cloud platforms offer fast infrastructure deployment and scaling, on-demand pricing, API control, and complete insight into what is running and where. A consequence of those qualities is that the cloud is capable of better utilization profiles than physical infrastructure. Unused servers can be discarded. Workloads can be balanced across servers to ensure optimal utilization.

At least that's the theory; in practice many organizations suffer from what's come to be known as cloud sprawl. It's easy to create a new server. Once a server has been deployed, it'll stay deployed unless someone intervenes to shut it down or it's shut down automatically. While it's running, the server will incur costs whether it's doing anything useful or not.

The problem arises when administrators don't implement proper cloud management and accountability processes. If a developer requests a test server, who shuts it down later? If extra servers are deployed to accommodate unexpected load, who will shut them down once the surge has passed? These are not technological questions, but management problems that can be solved by implementing processes to better account for and control cloud infrastructure deployments.