In the fixed scaling model, the same type of resource is always added to the service, such as additional web servers. This information can be a configuration parameter to the orchestration workflow that is set in advance by the administrator.

In the scale everything model, it is the application itself that scales out, and we require the location of the application in the service catalog from which it is derived.

With intelligent scaling, the localization process identifies the resource to be scaled and passes that information to the orchestration workflow to scale that component. Depending on the application requirements, the scaling workflow might be generic enough to address all components of the application as a single workflow responsible for the scaling execution.