In this context, the authors identify the
most common critical issues and constraints that have an impact on designing the
ETL processes. As the ETL design solution, the authors propose to apply a layered
infrastructure. The infrastructure is composed of typically used functionalities and
services in the ETL scenario and it is a basis for building various applications. The
infrastructure includes, among others: DBMS access modules, file access modules,
parallel read/write modules, and data processing modules. The authors also discuss
and give practical suggestions on implementation issues including database partitioning
options, parallel processing, and pipelining in the context of ETL.
xvi
Chapter V, Data Warehouse Refreshment, by Alkis Simitsis, Panos Vassiliadis,
Spiros Skiadopoulos, and Timos Sellis, focuses on methods for designing efficient
workflows of tasks within ETL processes. The features that make ETL challenging
include, among others: huge data volumes, assuring the quality of data, assuring
high performance, adapting workflows after changes in data sources, and changes
in a data warehouse itself. As a response to these challenges, the authors propose a
modeling approach/framework and its examplary application for the construction
of ETL workflows. This approach is based on the life cycle of the ETL processes.
The life cycle consists of four phases: reverse engineering and requirements collection;
logical design; tuning and physical design; and software construction.
Pages:
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35