MDM and Enterprise Architecture
digg this | 
del.icio.us | 
Reddit | 
Stumble It!
Editor’s Note: Another great guest post by Joan Lawson, a talented enterprise architect who worked for one of my clients in the software industry in 2003. For more information on Joan, please see her LinkedIn profile — Dan Power
Master Data Management (MDM) may be one more Three Letter Acronym (TLA), but it’s a central point in the practice of Enterprise Architecture. Together with SOA-based applications and a robust middleware platform, an ideal architecture is readily achievable.
Let’s take an example using party data including customers and prospects. Party data may have a “system of initial record” in any of the many ERP or CRM applications that a company may have.
A message with new party data can be written to the integration platform from the CRM application. Based on business rules, a Business Process Execution Language (BPEL) system can orchestrate the data management services in the MDM hub, write the clean party data into the MDM hub, and then message the clean data to the other ERP and CRM applications.
Ditto with product master data.
In this example, customer and product dimensions in the data warehouse are managed by the “source of truth” – the MDM hub. And the fact data for the warehouse (such as quotes, orders, and service events) can be sourced from the OLTP applications.
For those interested in real time monitoring of transactional data, consider placing that data on the integration platform as well. A Business Activity Monitoring (BAM) platform taps into that data to monitor it against KPIs. And once again, the MDM hub provides the “source of truth” for the master data.
The end result? Clean, consistent master data, whether used in the business applications, the data warehouse and business intelligence platform, or in real time business activity monitoring.
Please let us know by commenting here or on the MDM Community if you’re using MDM as part of your enterprise architecture.
Data Architecture (includes MDM) should be considered as a core part of Enterprise Architecture initiative and be hand-in-hand with any investments in integration architecture and SOA related technologies, if not it is not cost effective for organizations. (my opinions is that Enterprise Architecture shouldn’t exist without a comprehensive MDM strategy tagged with it)
Everything revolves around data, so the basic tenants and capabilities of Enterprise Data/Information Architecture should be in place prior to major business IT programs so that the new systems and processes are built in the “right” way. Planning to retrofit data architecture at later stages of deployment in large organizations will result in a recursive loop due to the inert nature of dependencies they/we create for data movement between these applications.
Gautam,
I understand what you are saying. Enterprise Architectures are key to the alignment of IT with business. An EA needs to include a comprehensive data architecture. I don’t believe that an EA shouldn’t exist without an MDM strategy. That being said, an EA must include the services of MDM – services for data management and access. Those may not be classified in an EA as an MDM strategy.
In response to your second paragraph, a high level Data/Information Architecture definitely needs to be formulated. New systems and processes need to be implemented in accordance with the new architecture – while existing systems and processes should be transitioned to the new architecture.
Joan,
I do agree in the principle that EA shouldn’t be tied to MDM strategy, but by my experience (that’s why prefixed it by stating “My Opinion”) they are very much intertwined for the success of EA. I have seen too much capital being spent on designing systems with little to no thought given to “single of source of truth” data. My reference to the inference of that statement was not just towards MDM strategy but towards the whole EIM/EDM strategy (Enterprise Data/Information Mgmt)
Thanks for the informative article