Also, the same report suggests considering ???one person per
source??? as a guide to accomplishing the ETL implementation effectively.
Based on the previous discussion, we can identify key factors underlying the main
problems of ETL workflows:
??? Vastness of the data volumes
??? Quality problems, since data are not always clean and have to be cleansed
??? Performance, since the whole process has to take place within a specific time
window and it is necessary to optimize its execution time
??? Evolution of the sources and the data warehouse can eventually lead to daily
maintenance operations
Visualizing and understanding this kind of system is another issue. In fact, traditional
modeling approaches need to be reconsidered: we need interactive, multiview
modeling frameworks that abstract the complexity of the system and provide
complementary views of the system??™s structure to the designer (apart from simply
providing the big picture, like the traditional ER/DFD approaches did). Moreover,
we need to be able to manage risk through our modeling artifacts. For example, we
would like to answer questions like:
??? Which attributes/tables are involved in the population of a certain attribute?
??? What part of the scenario is affected if we delete an attribute?
??? How good is the design of my ETL workflow?
??? Is variant A better than variant B?
Main.
Pages:
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244