This is the second critical aspect of any MDM tool. The first critical part can be found here.
On one side, MDM enforces data quality and builds each master data object. The data thus collected should be sent across the entire landscape system for all other dependent applications. This propagation task ensures that data is in sync between the MDM system and all other transactional systems.
MDM systems basically “pushes” updated data to all dependent systems within the application landscape. Alternatively, an application can also implement “pull” mechanism to read the updated version of master data at regular intervals. A DW/BI system will be one such application that will be in sync with updated master data from MDM system.
A propagation task is triggered when the workflow is completed. Data flow is always one-way between MDM and transactional systems. Typically, all dependent systems only keep a copy of the master data maintained in the MDM system and any changes to the master object has to be initiated via its respective workflow. Modifying the data and sending it back to MDM is avoided at all costs in order to maintain consistency.
The image above summarizes the functionality and data flow in a landscape that has implemented a Master Data Management system.