white rectangle.png
23
Amazon Elastic Compute Cloud
(Amazon EC2)
•Resizable compute capacity in the cloud
–Instances come online in a few minutes
•Accessible via web services API
•Instances look and feel like a remote commodity server
–1.7Ghz circa 2006 x86 processor
–1.75GB of RAM
–160GB local disk
–250Mb/s network bandwidth
•Pricing (pay for what you use):
–Compute: $0.10 / instance / hour
–Transfer: $0.18 / GB
•Instances also come in Large and X-Large sizes
–4x the base instance
–8x the base instance
–Linear pricing to base instance
The way the EC2 service is used is bascially linux boxes on demand.   Typical appl architecture are web servers, database servers, app servers—you would build up machines as you normally would on a linux box and then we provide tools to allow you to shrink-wrap your machine images, store them on S3.   The way Ec2 works is you can make a web service call and say I need 25 of those AMIs up and running—within 5-10 minutes you are up and running.   Under heavy load allocate more, if load drops off you can scale back.  You don’t pay for boxes sitting idle and you can handle any spike.

It’s a medium weight server (it’s an instance), hardware underneath are comodity servers.   EC2 uses Xen for virtualation layer.  Anythign that runs on Xen runs on EC2.