I have to model a deployment based on Docker containers. I see this answer which suggests to use nodes to model Docker containers.
I would prefer to use:
Device
for physical machines. It is the only element for physical resource, see 19.5.7.1Node
for virtual machines, which is compliant with the description provided by 19.5.10.1A Node is computational resource upon which artifacts may be deployed for execution
ExecutionEnvironment
for Docker containers. It is partially compliant with the description provided by 19.5.8.1An execution environment is a node that offers an execution environment for specific types of components that are deployed on it in the form of executable artifacts.". Partially because a container may host any types of artifact not specific ones.
ExecutionEnvironment
for web server, servlet containers or application servers. This is the usage shown in the specification. See fig 19.12Artifact
for binaries produced by our software forge.
The conclusion from of above is to separate artifacts such as WAR/EAR from Docker containers and to sperate physical or virtual machines from Docker containers.
Is my proposition logically and correct from an UML point of view?