MindTree technology stack consists of the out-of-the box data integration framework as well as various reusable components that encapsulate a variety of best practices to ensure scalability and flexibility while building complex DW/BI solutions.
At a logical level, thus RUBIC Technology Stack stands for a collection of best practices that help in making a data warehouse implementation more flexible, scaleable and nimble. A few examples of structured approaches that are encoded as a part of RUBIC, especially on the data warehousing area that is followed by MindTree are:
- Structured maintenance of integrated “Source data” to ensure ‘no loss of information’
- Structured reject management processes where each reject item is tagged with the reject reason and exposed to individual reject owners
- Automated incremental reject processing mechanisms to ease ongoing administration of the data warehouse
- Structured de-coupling of the various stages in the architecture to ensure scalability to changes
- Introduction of various report presentation best practices like contextual metadata sharing, KPI oriented charts presentation, alerts and notification mechanisms etc.
- Incorporation of some of the modeling best practices like ‘Generic Entity Modeling’ (enhances flexibility to changes) or ‘D-Leveling’ (supports storage of data at different levels of grain)
- Usage of performance aiding constructs as a part of the loading framework (Usage of database level features for areas requiring high levels of data processing as row-by-row processing in an ETL can at time cause performance constraints)
These logical approaches can be implemented with minor workarounds in any tool set.
