Analysis across an architecture model is possible through the propagation of
quantities through layers. A natural option for this is to consider workload measures
(e.g., arrival frequencies) that are imposed as a ‘demand’ on the model elements by
the users (located in the higher layers, e.g., customers). These quantities propagate
towards the lower layers, eventually being translated in demands on each model
element. Once workloads have been determined, we look at the effort these
workloads require from the resources (modelled by structural elements) and from
the behaviour elements (modelled by services, processes, and functions)