In a nutshell, at Amazon, and I assume other cloud hosts, you basically launch a server (EC2) based on your prefered OS (AMI), prefered hardware capabilities (instance type), region, zone, once launched connect and add whatever programs you need to the server (LAMP, etc), save your customized server as your own AMI for later use.
Then when you need to grow you can either expand out or up. Out being add additional instances behind a load balancer using your own AMI as a template or up by relaunching your current AMI using more hardware capabilities (CPU, RAM), ie a larger instance type. Both of these adjustments can be done in minutes either automatically or manually.
It is a powerful system that needs more than 20 minutes to fully comprehend and implement. This is probably why they allow new AWS accounts to launch what they call a micro instance for free for a year so people can play around with it without worrying about getting charged.
|