Master Data Management - Even more critical in light of Cloud Computing
Guest post by
Olivier Penel, Principal Consultant - MDM, Oracle practice, Enterprise Solutions, Infosys Technologies Ltd.
MDM (Master Data Management) initiatives should be seen as a way to provide master data 'as a service' across the organization, lines of businesses, channels, countries, etc. The idea is to decouple the management of core data entities (products, customers, locations, etc.) from the business applications (ERPs, CRM, range planning, merchandising, supply chain, e-commerce systems) and to provide those systems with the data they need, at the right time, and in the right context.
The SOA (Service Oriented Architecture) layer aggregates master data from MDM systems, but also transactional data, historical data, or reference data, from internal and external data sources.
For example, organizations often subscribe to data pools (e.g. GDSN) as a trusted source of information to enrich their master data repositories. Large manufacturing companies have long been in that space, making their product information available in an automated and standardized manner. However, small and medium suppliers cannot really afford investing into PIM (Product Information Management) solutions and GDSN (Global Data Synchronization) infrastructures. To deal with this issue, GS1 UK recently launched a new initiative intended to provide those suppliers with a "PIM as a Service" offering that will eventually provide retailers with trusted product data from a much bigger number of their suppliers. This move clearly shows that the market is fast maturing to the idea of sourcing master data from external sources in an automated and standardized way.
We can see very clearly that the boundaries between the enterprise information systems and the outside world (partners, suppliers, customers, data pools, etc...) is not as clear as it used to be and the current move towards Cloud Computing is likely to increase this pattern, especially in the MDM area. In this context, one can wonder what the trend towards Cloud Computing mean for organizations that are looking to launch an MDM initiative or to extend their existing MDM solution.
I believe that Cloud Computing is very likely to increase the need for MDM and strong Data Governance. The main reason for that is the obvious risk associated to Cloud Computing, which is a loss of control over data and data quality. MDM is a critical step towards strengthening this area and mitigate the associated risks.
The trend towards applications and data in the cloud will require more sophisticated systems and strong data governance processes and technology. It is delusive to think that good enterprise wide services can be set-up without underlying data governance. Who owns data? Who can access or change those data? What data quality means for the organization and how to monitor it? Those are just a few questions that an organization should answer before starting an SOA or Cloud Computing initiative.
The fracturing of data caused by the adoption of Cloud Computing raises the importance of MDM in keeping disparate data synchronized. Cloud Computing will become extremely confusing and totally inefficient without a consolidated version of the truth in a proper master data management hub. In addition, as applications go off premises, there is a real danger MDM could get out of reach very quickly. Therefore, MDM should be seen as a pre-requisite to cloud computing, in order to get control over data before trying to leverage external data sources, or external applications consuming master data.
On the other hand, Cloud Computing will introduce new opportunities with using external reference data pools (GS1, GXS, SA2, 1SYNC, Dun & Bradstreet, etc.) or data quality services such as identity and entity resolution. As a conclusion, organizations looking at Cloud Computing as a strategic way forward should consider:
- Using MDM to create a single version of the truth before enabling cloud-based applications
- Combining internal master data with external cloud-based data sources to create a consolidated version of the truth using a co-existence MDM pattern.
Comments
Olivier,
This is a great article. It is highly relevant for product data domain. However specific to customer domain, would a client be willing to share customer data outside its DMZ zone. Preferably not.
Great article
Jairaj
Posted by: Jairaj | November 19, 2010 11:01 AM
Not many companies can afford an MDM investment in such cases cloud services could precede MDM. In other words a cloud service offering Marketing information to a company looking for a campaign management is a viable cloud service customer. One necessarily dont have to invest in MDM from this perspective. In other words Cloud and SOA are pretty going parallel with this view.
Posted by: Satish Peri | December 6, 2010 11:05 AM
Good Article.
While the business processes and related transactions be serviced through the externally hosted servcies / clouds it becomes the need of the system to serve customer , product and other master data domains to be available through the cloud as well. Yes the challenges of security , DMZ all have to be tackled. MDM implementations enabling improved data quality services , business services and SOA based integration architectures, shiping MDM also on to cloud would be feasible. Lot of cleanup and preperation work required and MDM initiatives would help with quality cloud deployment.
Posted by: Rajasekar L | January 14, 2011 11:28 AM