This is the ability to scale up and down services rapidly for on-demand computing. E.g. be able to add 20/100/500 instances to a service in 5 minutes.