BI TRAININGS:

Course: #BI-2D - DW Hybrid: key to significantly reduce your EDW TCO (2 days).

A little dated (before the era of Big Data) but the course covers all the required notions and remains relevant since it describes what the consumption zone of ​​a Lake House should look like. Click here for a description of the LakeHouse architecture (coming soon)

We have created a hybrid DW architecture (ITA&S-H) to essentially:

  • Reduce the amount of data movement from sources to EDW to data marts
  • Increase agility
  • Reduce costs (the reduction in total cost of ownership comes from the drastic reduction in the number of tables, which translates into a major reduction in the amount of ETL and therefore costs)
  • Be able to build some of the data warehouses in near real time (real time is simplified as there is only one set of ETL/ELT processes to run)

The ITA&S-H architecture resembles a constellation of snowflakes with the addition of conventional normalized data structures (called EDW extensions). It is therefore a hybrid of the EDW and Datamarts architecture merged into a single database model with a series of improvements such as:

  • Fact adjustment records via reversal record and new record
  • Conformed Dimensions with the principle of the lowest level
  • Multiple anchor points in one dimension
  • Automated Dimensional Recasting
  • Denormalized Snowflakes

We have successfully used the ITA&S-H approach:

  • For traditional BI projects (multiple times)
  • For Data Lakes on Hadoop (multiple times)
  • For Lake Houses on Azure (several times)