New Data Warehouse Technologies - Data Warehouse Systems: Design and Implementation - page 523

a

Data
Sources

Staging
Database

Staging
Database

Data
Warehouse

Extract Transform Load

b

Data
Sources

Staging
Database

Data
Warehouse

Data
Warehouse

Extract Load Transform

Fig. 13.7 ( a ) Extraction, transformation, and loading (ETL) process. ( b )Extrac-
tion, loading, and transformation (ELT) process

data warehousing. Further, the amount of operational data produced daily is
constantly increasing due, among other reasons, to business globalization and
the explosion of the number of transactions over the web. In this scenario,
it is likely that the time needed to refresh the data warehouse using the
traditional ETL process exceeds the allocated updating window.
The above discussion aims at explaining why some practitioners and
vendors are proposing a different data loading paradigm: the extraction,
loading, and transformation (ELT) process. We discuss this next.
Consider Fig. 13.7 , which provides a detailed look of the data staging phase
in the back-end tier of the architecture depicted in Fig. 3.5 . The figure shows
that during the ETL process, data are loaded from the sources into a staging
database, where the necessary data transformations occur, as described in
Chap. 8 . After this process, the transformed data are loaded into the data
warehouse. The process guarantees that only data relevant to the solution will
be extracted and processed, potentially reducing development, extraction,
and processing overhead. This also, in some sense, simplifies the management
of data security and therefore the data administration overhead. On the other
hand, accounting just for relevant data implies that any future requirements
that may need data not included in the original design will need to be added to
the ETL routines. This may lead to important redevelopment tasks. Besides,
the use of third-party tools to implement ETL processes requires learning of
new scripting languages and processes.