Lots of people proclaim the performance advantages of cloud computing. These folks, I think, are mostly purveyors of cloud computing services. Which does not mean that cloud computing has no advantages or offers no improvements in performance. But it also does not mean that all improvements from migrating to the cloud are derived from the cloud.
Yes, cloud computing can reduce administration costs, mostly by standardizing the instances of hosts to a simple set of virtualized machines.
And yes, cloud computing can reduce the infrastructure costs of servers, since the cloud provider leverages economies of scale (and virtualized servers).
But a fair amount of the performance improvement of cloud computing comes from the re-architecturing of applications. Changing one's applications from monolithic, one-program-does-it-all to smaller collaborative apps working with common data stores and message queues has an affect on performance. Shifting from object-oriented programming to the immutable-object programming needed for cloud computing also improves performance.
Keep in mind that these architectural changes can be done with your current infrastructure -- you don't need cloud to make them.
You can re-architect your applications (no small task, I will admit) and use them in your current environment (adding data stores and message queues) and get those same improvements in performance. Not all of the improvements from moving to a cloud infrastructure, but the portion that arises from the collaborative architecture.
And such a move would prepare your applications to move to a cloud infrastructure.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment