End-to-end Data Flow between MDG and ERP systems.

In this blog we will be discussing data; the various types of data, its flow between MDG hub system and ERP system. Before we start, let's list the types of data we will be discussing here: Master data, Transactional data and Customizing data.
Master data is the only type of data end users (data stewards) will be modifying within MDG system. Material is an example of master data. It includes the identification data (e.g. Material ID), descriptive data ( e.g. material description) and process data(e.g. material type). Changing the value of material description does not affect the material unit of measure or characteristics.

MDG maintains master data while ERP uses the master data.

Arrow 1:
Lets consider the specific case of "Material" as the master data.
Customizing data is created by MM functional person. This customizing data created (as a result of configurations) by MM functional person, influence the way MM related end user (like MM01, MM02 etc) transactions in ERP, behave at runtime. e.g. Field control settings.
The same customizing settings influence the way MDG screens work as well. In other words, MDG uses a lot of this customizing data for its functioning. For instance, Field validations: MDG 'Create Material' screen would have the same field validations as MM01 screen on ERP. The field control settings, in both cases, are based on MM customizing settings.

If MDG is implemented as a Hub system (as opposed to being co-located on ERP client as a true add-on), it does not immediately have access to the customizing data in ERP. In this case, the customizing data has to be explicitly moved from ERP system to the MDG system. MM functional person can identify the specific data and tables that needs to be copied. This document provides a great overview of the alternate strategies used for customizing synchronization.

Arrow 2:
In addition to creating master data, MDG is used to perform updates to existing data.
Existing data is also used, in Embedded Search scenario, to search and ensure that a duplicate master data entry does not already exist.
Where does the existing data in MDG come from?
This is created, at system go-live/cutover, by an initial data load. All the master data (cleansed version) is moved from client systems to MDG system. in case of MDG-M, the data is loaded to MARA, MARC etc tables.

Arrow 3:
Once the MDG solution is implemented and active, master data specialists will use the MDG functionality to create master data.
This data will proceed through the approval workflow. Once approved, the master data gets 'Activated'. At this point, data moves from MDG staging tables to active tables (in the case of ReUse).  If MDG were installed as a standalone hub, the data is still in ERP tables in the MDG system. It has to be replicated to the corresponding master data tables in the ERP client system.
This is implemented using the Data Replication Framework (DRF).

Please note that I assume ERP to be the client system. In generic terms, this down stream system/ that gets master data from MDG can be any system (Legacy, ERP, SRM, CRM etc).

Images: MM customizing transactions in SPRO, MDG customizing transactions; generated MDG tables; Data in MDG tables; Data in Mara; SMT mapping