Master Data Management Comes of Age
By David Butler on Nov 01, 2011
MDM is on a role. There was overwhelming interest in MDM at OpenWorld last month. Literally thousands of people attended our MDM sessions and observed our MDM Value Chain capabilities at the MDM demo-grounds. MDM interest was also exemplified by our 5th consecutive 50% YoY growth quarter. These successes lead me to reflect on how MDM has come of age and what lead me to it in the first place.
After cutting my teeth on batch and time-sharing operating system development in the 70s, I got in on the ground floor creating the world’s first database computer (the Teradata DBC1012). Data Warehousing was born. The 1012 stood for 10 to the 12th power, or one terabyte. That was a big number in the 1980s. Today, almost every IT organization in the world manages a data warehouse and one or more data marts. My work creating the Teradata database and watching how customers used it, taught me how powerful analytical models could be, but it also taught me how limited they were in actually achieving a single view of what is going on in any heterogeneous environment. The root problem was that data is created inside operational applications with transactional data models while data warehouses were built on star schemas. The two could not be brought together without an engine that can support mixed workloads and a data model that could keep up with transactional applications as well as provide decision support. If I was going to be a part of creating that kind of infrastructure, I had to move to Oracle.
At Oracle, we had a parallel processing platform called Oracle Parallel Server. It was the foundation we used to create Real Application Clusters (RAC). This represented the platform needed to solve the single view problem. It could do mixed workloads with ease and had the benchmarks to prove it. Today, RAC is the foundation technology supporting a truly remarkable next-gen database computer – Exadata. It does what the Teradata system cannot do – namely support transaction rates under operational applications at the same time that it runs data warehouse analytical workloads. This was a foundational breakthrough, but we were still left with the data model issue.
Without a data model that can support operational applications across all lines of business, we could not create the enterprise wide authoritative dimensions needed to unify operations and analytics. I ran into this data model at Oracle quite by accident. In the late 90s, I was moved to application development in the Communications vertical. I was charged with developing a billing system that would leverage RAC. At the same time, Oracle was going through one of its great leaps forward by consolidating all applications into one suite. I had to provide requirements for the data model under this suite to insure that it supported a single view of products and customers for the Comms industry in general and my new billing system in particular. This is what I was looking for: an item master and party model that could support operational applications and provide most of the dimensions needed for the star schema! In 2001, the E-Business Suite was released, and this data model began to prove itself across industries, lines of business and geographies.
To my mind, MDM was born with this release. A lot of work remained to make it functional in a heterogeneous application landscape. Items such as source system management, data governance workbenches, hierarchy management, data quality tools, application integration and business intelligence mappings have been added over the years. But it was the combination of mixed workload support and a robust transactional model that made the unthinkable possible.
Today, organizations in every industry are realizing significant measurable ROI, reducing risk, bringing corporate governance to master data, increasing compliance at lower costs, getting more out of existing systems, speeding M&A activity, reducing spend, increasing customer loyalty, accelerating new product introductions, optimizing business processes, operationalizing data warehouses, increasing the accuracy of enterprise reporting systems, supporting service oriented architecture deployments, consolidating ERP and CRM systems, and rationalizing the overall applications landscape.
It is not very often that a technology comes along that can measurably assist organizations across such a wide variety of top business and IT initiatives. Master Data Management is one of these rare breeds. It is finding its way into every aspect of IT operations facilitating the flexibility required for full IT and business alignment. In the near future, it will be hard to find an organization with an information architecture that doesn’t have one or more MDM hubs included. This is the hallmark characteristic of a technology that has indeed come of age.
For more information on this topic, don’t miss my new article Master of the Data Universe in the November edition of Oracle Profit Magazine. And if you missed OpenWorld, you can find all the MDM presentations in the Content Catalog. Just select the Master Data Management track and click ‘search’.