The top-level Package Diagram (PD) in each UML model includes four packages, one for each of the views in the model. One of these packages, called processors, may be decomposed through one or more further levels of PD to a set of Deployment Diagrams (DPDs).
DPDs are UML’s attempt to recognise that there is a world outside software. That is, these diagrams attempt to show the physical processors in the system (called nodes in DPDs), the processes that are allocated to them, and, to a limited extent, how these processors are interconnected.
These diagrams are intended to show the allocation of processes to nodes, by listing the processes that may run on a processor underneath the node symbol for the processor. Cradle will support this mapping in a different way, by providing a frame called PROCESSES in the node’s component specification, in which the names of the processes allocated to the node can be listed.
By being defined by component specifications, the descriptions of nodes will be cross referenceable.
DPDs do not show data exchanges between nodes. For a more complete representation of architectures, consider Architecture Interconnect Diagrams (AIDs) and Process Architecture Diagrams (PADs).
DPDs are not hierarchical. Their connectivity is:
DPDs are only available in models in the Implementation Domain.
DPDs take their name and number from the package from which they are an expansion. They contain nodes and links. The nodes are defined by specifications. The number of the specification is the dot concatenation of the DPD number and the node number, its name is the same as the node name.
An example DPD is:
The symbols available in DPDs are: