When you have a dashed arrow with the <<deploy>>
keyword, one end will be a Node
and the other end will be a DeployedArtifact
. This indicates that the Node
can support the DeployedArtifact
and is an alternate representation of putting the DeployedArtifact
symbol inside of the Node
symbol.
Here is a graphical representation of the two options, from page 659 of the UML 2.5.1 specification:
In both cases, the artifacts have been deployed onto the node.
I think that it would be a safe assumption that if an artifact has been deployed to a node, then it is being used on the node. If it's an executable, then the node will, at some point or under certain conditions, execute it. If it's a configuration file, then the node will, at some point, read it and use it.
Typically, I've shown multiple executions of a specific artifact in the textual description around the UML model. There may be other ways of doing it, such as multiplicity on the <<deploy>>
relationship or including multiple instances of the artifact within the node, but I've found that these get confusing to readers. Simply showing that an artifact lives on a node and providing other text or tables around the number of instances is more straightforward.