I'm facing a modeling question in MEGA Architecture regarding : Referential data and applications "owners" of them.
In our landscape, some data, named as referential data, are owned by specific applications and used (and/or) copied by other application.
For example: CRM owns Customer referential data, they are used by our production application. Production application owns also , let says, product referential data.
Each referential data is described by a Data Model (DM).
From here it's ok.
We know : in application properties, complement tab, we can link DM to the application. In our little example, Production application is linked to Customer referential data (because it needs them) and product referential data.
This part is also ok.
But how model the fact that production app is the owner of product referential data ? Same for the CRM ?
The idea behind is to make an impact analysis, if for example, we would like to change the CRM tool ...
Second point : how to model the ownership of referential data in an "application environnement diagram" ? As we can't place DM graphical object into the application graphical box ?
Could we use : database graphicla diagram ? Or just add texts in application boxes ?
Thanks a lot for your advices.
Solved! Go to Solution.
We are usually using the object "Database" because it can be a logical object too (we add a blue color cylinder in the list of mgs forms, same color as Entity MD.
So, Entity (MD) can be connected to application that is equivalent to : the application is owning or using the referential data ...
In an application diagram , you can put a database object (one per referential data or one for a small amount of data to group it in a "business" way ...)
You can link Entity (MD) / referential data to it ( a referential data will be link to only one DB). A compliance rule can be done for that
Then you know that a referential data owned by a "logical" database owned by an application is owned by this application.
In an IT city plan diagram, the same way because you can place both DB and Entity (MD) objects that is signifying that the data is owned and/ or used by the functions/services of the IT city plan area ...
In addition we are keeping the direct link app/entity (MD)
That's a way to do it with a pivotal object, the "functional database" and no metamodel extension.
Please note that sometimes we are making a new link between databases to associate functional and physical/applicative one ...
Thanks for your reply.
The DB Object is what we used (our first approach). You are right, the functional DB with another graphical color, could be done in this case without any change of the MEGA meta-model. Also we need to make a difference with physical DB through properties (maybe ?).
We'll do like that in next steps and see if it's relevant for our context.
That's right, it's why we are very often making a metamodel extension in order to have a computed form based on the value of type in a database object : functional, physical ...