ITIL Capacity Management encourages us to carry out capacity management at three levels, business, service and component. This is sound advice and I've carried it out regularly and found that the amount of effort required to model and capacity plan a service can be defined by the number of individual entities in each of these ‘layers’ of capacity management.
The purpose behind this layering is that capacity management should be done across the enterprise; from the business (who represent the user community and drive changes) to the technical staff who ensure that there is sufficient hardware in place. It is a sign of process maturity to be able to discuss capacity management in business terms (to include users) but also necessary to do so in technical terms (because it is necessary to have the right amount of hardware).
So what about the service layer? If there is a direct relationship between business volumes and component consumption then considering the service layer just becomes an overhead. If there is no such relationship then considering the service layer may provide the appropriate workload characterisation in order to model effectively. When this is true then I have found that the terminology in the service layer is typically understood by the business and therefore there is no need to consider the business layer.
If a direct relationship exists between the business layer and the component layer then use it. If there is a direct relationship between the service layer and the component layer that is also understood by the business then use that. It can reduce the workload by a considerable amount.
