System-Subsystem Horizontal Transition

Hi All,

I have a setup whereby I am using the System to Subsystem transition add-on to vertically transition two logical components; sub-system 1 (SS1) and sub-system 2 (SS2) to the SA layer of their own respective models.
Where SS1 has derived functions i.e. those not decomposed from the system model, that also impact SS2, I want to make sure our process is robust such that those functions allocated to SS2 in the SS1 model can be pushed to the SS2 model.
I believe this can be achieved using a horizontal transition between the SS1 and SS2 models, but I’m having a some issues:

  1. Although the SS1 component in the SS1 model and SS2 model have the same SIDs, the diff/merge window does not seem to relate them and instead creates a new “System” component
  2. The diff/merge tool warns of duplicate match keys (which I’m reasonably confident relate to the SS1 and SS1:SS1 elements, but I don’t know how to confirm this)

Does anybody know how to force the diff/merge tool to relate the SS1 component in the SS1 model with SS1 component in the SS2 model?
Does anybody know how to resolve the duplicate match key issue?

Please see the images below to help explain my situation.

In the short term I’ve discovered it is possible to transition the functions and not the allocations and do the allocation in the resulting model, but it would be nice to have this done automatically by the tool.

I’m using:
Capella 6.1.0
System to SubSystem 1.6.2

Many Thanks,


Duplicate match keys

I suspect you are breaking the philosophy of the transition approach and the intent of functional decomposition. The logical components “Sub-System-1” and “Sub-System-2” should be complete at the point of transition. New functions and components need to be hierarchically and structurally decomposed. By adding a new level-one function to the derived system you are breaking this completeness. The “Derived SS2 Function” needs to be introduced into the parent model’s logical architecture and then re-transitioned.

I am a bit suspicious of the duplicate key issue. I would have thought that all the elements should have unique GUIDs that are associated during the transition. This can be confirmed by inspecting the raw (XML) model data.