If a Component Exchange (CE) is created in System Analysis (let’s say between an actor and the “System”) it is transitioned into the Logical Architecture as a CE between a transitioned actor and the top level “Logical System”. That is fine.
However, after performing a transition to Physical Architecture, this top level Logical System is not transitioned (I checked that the “Propagate logical component breakdown in physical architecture” is selected).
Therefore, the initial CE is lost and only the actor’s port is present in the Physical Architecture.
Similarly, in the examples, I wasn’t able to find CEs that were created in System Analysis transitioned into the Physical Architecture.
Am I doing something wrong or the best practice would be not to create CEs in System Analysis?
Thanks for any feedback.
It is perfectly allowed to create CE in System Analysis and to transition them to Logical and Physical Architecture.
For your problem, I think the CE is not lost in Physical Architecture, but connected to the Physical System (which can’t be displayed on architecture diagram). Thus you should be able to retrieve your CE by moving manually the port from the Physical System to another Behavior Component (drag & drop of the port from the Model explorer).
In a more generic case, I think your CE should not be connected to your Logical System. Your Logical system should include a first architecture and thus several ‘sub’ logical components, your CE should then be moved to one of the sub-component. Then during the transition Logical-Physical, a new Behavior Component will be created, with all its CE.
So I would recommend to avoid having CE on your Logical System when you launch the transition, it will be far easier to manage, and you will have a first architecture at Logical level !