Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model promotes availability and is composed of five essential characteristics, three service models, and four deployment models.

On-demand self-service.
Broad network access.
Resource pooling.
Rapid elasticity.
Measured Service.

Virtualization is not required to meet any of those, you can rapidly provision physical nodes in a number of ways. Most clouds copy images across the network to the VM host machine when you launch a new image, there is no difference to copy over an OS that runs without virtualization. It is also possible to PXE boot and have a diskless host machine without virtualization if you want more rapid boot times.

Cloud computing did exist prior to virtualization, we just didn't call it “cloud” and most people didn't use it because of the price of resources — many enterprises have been running multi-tenant application grids for a decade or more that meet the definition above.

Now thanks to globally ubiquitous network access public cloud computing is interesting as well. Combine that with Moore's law and its ability to continually decrease the price per compute unit and on demand computing and data analysis job processing is interesting to more and more folks.

This post was mentioned on Twitter by bpiatt: New post: How to tell the difference between “cloud” and “virtualization” http://www.bretpiatt.com/blog/?p=112...