Traditionally we have been doing ETL; extract of the data from the source systems, then transforming it to fit our needs, and finally loading them into the end target such as a fact table. 

However, in the world of Massive Parallel Processing (MPP) scale out data warehousing, can this approach still be used?

Some advocate the ELT approach instead, which extract the source data, then load it into the end target, and finally do the transformations at the data in the target. But what are the benefits of this? And does it really matter?

This session tries to answer these questions by looking at the difference between the ETL and ELT approach, and how SSIS can be used with a Parallel Data Warehouse.

(no tags)