Effectiveness of Data Vault compared to Dimensional Data Marts on Overall Performance of a Data Warehouse System
The purpose of this pilot study was to compare the traditional Kimballs approach (using star schemas Dimensional Model, directly from the data sources source) to define Data Marts vs. newest Linstedts approach (using Data Vault 2.0 modeling of EDW with virtual/materialized Data Marts as star schemas). The comparison is performed experimentally using a specifically designed, realistic and broadly applicable case study. The study includes implementing a Customer Experience Data Mart in the telecommunication domain extending the work done by the TM Forums standards/guidelines demonstrating a feasibility of a complete implementation. A generic application implementation is completed with experimental test data and our comparative analysis included both qualitative and quantitative elements.
The paper also presents a novel systematic technique for data vault modeling, in the presence of highly generalized source entities, emerged from this experiment. This satellite per sub-type technique is devised to fit any highly generalized industrial source data model.
Keywords: Data Warehouse (DW), Data Vault (DV), Extract Load Transform (ETL), Data Mart (DM), Raw DV, Dimensional Model, Hub, Satellite, Link
Download Full-Text
ABOUT THE AUTHOR
Zaineb Naamane
Computer Science Department, Georgia Southern University
Zaineb Naamane
Computer Science Department, Georgia Southern University