Page 53 -
P. 53

14     Enterprise Data Governance


                                 Unification of the repository

                                 MDM brings with it a unified and centralized management
                                 of reference and master data. This does not mean that the
                                 physical architecture must be centralized. It is possible to
                                 opt for a distributed technical architecture.  We will return
                                 to this IT aspect in Chapter 12.



                                1.3.1.  Data initialized before use by transactional
                                systems

                                  Each data  initialized  before its  use by the transactions
                                that support the company processes, is a reference or master
                                data. This data is often administered by heterogenous,
                                technical and sometimes even simple automated office tools
                                (such as  spreadsheets). They  are employed by teams
                                responsible for the functional and IT parameter settings.
                                  The structure of this data can be simple (this is often the
                                case with codes and labels).  This is reference data. More
                                complex data is also involved, associated with  validation
                                rules and integrity constraints. This is the case for instance
                                with product configurations, organization definitions, charts
                                of accounts, etc. This is master data.

                                  For these types of data, an MDM system  replaces the
                                existing technical and heterogenous tools. It brings a unified
                                business  data governance, across the whole of the
                                Information System, with high value added functions such
                                as version  management, data input by use contexts, data
                                traceability, etc.


                                1.3.2. Duplicate data

                                  When a data element  is duplicated in multiple systems,
                                update synchronizations must be dealt with. The data must
                                have the same value in all of its storage zones. In order to
   48   49   50   51   52   53   54   55   56   57   58