TABLE OF CONTENTS This section introduces deployment diagrams, which realise a structural aspect of the model of the system. Like component diagrams, deployment diagrams are often hailed as being used towards the end of a project life cycle when the system has been packaged up, although this is often not the case. If a component diagram shows 'what's in the box', it is the deployment diagram that shows 'where to put the pieces'.
Deployment diagrams are made up of two main elements: nodes and relationships. Nodes represent real-world aspects of a system and show where artefacts from the component diagram are housed. Nodes are represented graphically by a 3-D box, on which instances or artefacts are shown.
Figure 5.78 shows the partial meta-model for deployment diagrams in the UML. It can be seen from the diagram that a 'Deployment diagram' is made up of one or more 'Node', one or more 'Relationship' and zero or more 'Deployment spec'. There are three types of 'relationship': 'Association', 'Generalisation' and 'Dependency'. The 'Association' and 'Generalisation' relationships are the same as for all structural modelling but the 'Dependency' relationship has two specialisations defined - 'Manifestation' and 'Deployment'. The 'Manifestation' relationship describes the relationship between a component and its associated artefacts so it may be thought of as a type of realisation. The 'Deployment' relationship describes an artefact being deployed onto a particular node. Such a deployment may have a 'Deployment specification' associated with it that describes features of the...
