MDM Performance Tuning and PlanningMDM Performance Challenge |
 |
Master Data Management (MDM) software solutions solve mission-critical business issues dealing with the trustworthiness of data that is essential for customer service, sales, marketing, and regulatory compliance. Customer and product data from across the organization is consolidated, compared, matched and merged to create a trusted source.
This process also creates a large repository of source data, merged data, relationship data, and meta-data that makes MDM solutions vulnerable to performance issues. With large and growing volumes of data there are many “performance points” that are at risk for sub-optimal levels of individual performance and when combined deliver unacceptable levels of MDM system performance.
Examples of common business issues due to poor MDM system performance are:
- Degradation of operational performance due to long “batch windows” for processing reference data.
- Delay in real-time activity when a transaction relies upon a poor performing MDM system for information.
- Frustration and lack of willingness to use MDM Systems with slow response times by both Data Stewards and other users of the MDM system.
- Rapid decrease in performance as the amount of data increases on already slow systems.
Tuning – The MDM Performance Solution
While complex, an MDM system with a 3-tier architecture can be tuned to improve performance. Each tier has its own resource dependencies, and “Performance Points” — locations where performance issues are caused. Using our MDM experience, the GlobalSoft team has built a performance review and testing methodology that quickly determines bottlenecks slowing your system down and identifies the best options to remove them to improve your MDM system performance.
The Approach to Improve MDM System Performance
Performance requirements vary for each MDM deployment. GlobalSoft works on projects from planning stages for new systems to production environments exhibiting slow performance. GlobalSoft uses a multi-faceted approach to Performance Tuning and Planning. Each aspect provides different insight into the root cause and the likely improvement in performance from tuning.
Performance Review
A qualitative assessment of performance is important to planning for future growth or a new phase — ensuring future development will accommodate performance needs.
During the review, the team collects current performance data, environment details and MDM configuration. This data drives a report of recommendations and expected performance based on best practices and GlobalSoft’s experience. This will include observations around DB parameters and infrastructure as well as MDM configurations.
Performance Characterization
Characterization is a process to quantify the impact of a set of recommended changes that came from a performance review. A series of test runs are made based on individual changes measuring the actual performance impact of the adjustment. Based on the results, the most promising combinations of changes are selected for additional testing. The data is then consolidated into a report showing the projected performance benefits of each potential choice.
Performance Benchmarking
If an MDM environment is experiencing performance issues, Benchmarking is an expedient way to determine where to focus efforts. A suite of benchmark tests are run on a copy of an existing ORS.
These tests are run on a hardware test setup with known performance characteristics. These tests can quickly determine if performance characteristics are driven primarily by underlying infrastructure or by hub configuration.
