CDI vs. MDM
CDI is a subset of MDM.
Master Data Management (MDM), also known as Reference Data Management, focuses on the management of reference or master data that is shared across applications, IT systems, geographies or other groups. MDM is required for consistency of data and computing results, especially where the data resides in multiple systems. Customer data is one of the more important types of data that needs to remain consistent. But the needs in CDI are different from most other areas of MDM.
Frequently, MDM disciplines are called into play in large companies, which have different IT systems in various divisions or business functions (e.g., finance, sales, HR, engineering, etc.). These diverse systems usually need to share key data that is relevant to the parent company (e.g., customers, products, and partners).
MDM is also a critical IT function when companies exchange data with one another. B2B communications such as inter-bank transactions, orders, and routing instructions all need to cross corporate boundaries. Emerging standards, particular the various XML dialects for intercompany communication, have greatly facilitated this kind of messaging. MDM encompasses communcations standards, process standards (e.g. straight through processing for T+1 equities trade settlement in finance), and reference data management (e.g. security master file, cross-referencing cusips, SEDOLs, ISINs and other security identifiers).
Many computing models differentiate between three types of IT components: OLTP transactional computing (typically ERP), DSS (Decision Support Systems) and MDM. MDM is not only required for interoperability between different ERP systems, but also to supply meta-data for aggregating and integrating transactional data. This use of MDM is also necessary for Data Warehouse projects typically incorporated in Decision Support Systems. Since MDM typically is employed within the larger context of an enterprise architecture, MDM systems often provide a meta-data abstraction layer. This design allows systems that use the master data to address (call, publish, subscribe, etc.) the data as an entity relationship (ER)-scheme, which facilitates data management and connectivity.
CDI, therefore, is that subset of MDM focused exclusively on the customer record.
I think part of the confusion arises from the fundmentally different toolsets that are required to master customer data. Most reference data is relatively static; it grows obsolete and outdated and needs to be kept fresh with new data additions, but the parts numbers in inventory don't change very often. Customer data, on the other hand, degrades at a rate of about 2% per month, and CDI systems are much more focused on managing conflicting sources of data and arbitrating betweens source systems.
Master Data Management (MDM), also known as Reference Data Management, focuses on the management of reference or master data that is shared across applications, IT systems, geographies or other groups. MDM is required for consistency of data and computing results, especially where the data resides in multiple systems. Customer data is one of the more important types of data that needs to remain consistent. But the needs in CDI are different from most other areas of MDM.
Frequently, MDM disciplines are called into play in large companies, which have different IT systems in various divisions or business functions (e.g., finance, sales, HR, engineering, etc.). These diverse systems usually need to share key data that is relevant to the parent company (e.g., customers, products, and partners).
MDM is also a critical IT function when companies exchange data with one another. B2B communications such as inter-bank transactions, orders, and routing instructions all need to cross corporate boundaries. Emerging standards, particular the various XML dialects for intercompany communication, have greatly facilitated this kind of messaging. MDM encompasses communcations standards, process standards (e.g. straight through processing for T+1 equities trade settlement in finance), and reference data management (e.g. security master file, cross-referencing cusips, SEDOLs, ISINs and other security identifiers).
Many computing models differentiate between three types of IT components: OLTP transactional computing (typically ERP), DSS (Decision Support Systems) and MDM. MDM is not only required for interoperability between different ERP systems, but also to supply meta-data for aggregating and integrating transactional data. This use of MDM is also necessary for Data Warehouse projects typically incorporated in Decision Support Systems. Since MDM typically is employed within the larger context of an enterprise architecture, MDM systems often provide a meta-data abstraction layer. This design allows systems that use the master data to address (call, publish, subscribe, etc.) the data as an entity relationship (ER)-scheme, which facilitates data management and connectivity.
CDI, therefore, is that subset of MDM focused exclusively on the customer record.
I think part of the confusion arises from the fundmentally different toolsets that are required to master customer data. Most reference data is relatively static; it grows obsolete and outdated and needs to be kept fresh with new data additions, but the parts numbers in inventory don't change very often. Customer data, on the other hand, degrades at a rate of about 2% per month, and CDI systems are much more focused on managing conflicting sources of data and arbitrating betweens source systems.
2 Comments:
Interesting. I thought MDM was primarily SAP's term for CDI. I guess this means companies like Goldensource and Asset Control are technically in the MDM space as well.
Orchestra Networks has a blog on Master Data Management that has some good in-depth articles on the topic.
Post a Comment
<< Home