Next: Parameter Model
Up: Usage Model
Previous: Structure
Contents
Index
The PCM usage model has been designed based on meta models such as the performance domain model of the UML SPT profile [36], the Core Scenario Model (CSM) [37], and KLAPER [38]. It is furthermore related to usage models used in statistical testing [39]. Although the concepts included in the PCM usage model are quite similar to the modelling capabilities of the UML SPT profile, there are some subtle differences:
- The usage model is aligned with the role of the domain expert, and uses only concepts known to this role. It is a domain specific language, whereas the UML SPT performance domain model is a general purpose language that includes information, which is usually spread over multiple developer roles such as the component assembler and the system deployer, so that a domain expert without a technical background could not specify an instance of it. Nevertheless, domain experts should be able to create PCM usage models with appropriate tools independently from other developer roles, because such models only contain concepts known to them.
- The number of loop iterations is not bound to a constant value, but can be specified as a random variable.
- The control flow constructs are arranged in a hierarchical fashion to enable easy analyses.
- Users are restricted to non-concurrent behaviour, as it is assumed, that users only execute the services of a system one at a time.
- System service invocations can be enhanced with characterisations of parameters values, as described in Section 3.5.3.
Next: Parameter Model
Up: Usage Model
Previous: Structure
Contents
Index
Snowball
2007-03-16