Databases Reference
In-Depth Information
The most successful implementations of the top-down data architecture are
grounded in business requirements. The overall approach is iterative. With
each iteration, the scope of data is set based upon business objectives and
requirements. Then work begins on the data. The data sources are identified
and the data is extracted. Each layer of the data architecture is defined,
starting with the source and moving forward, designing each step along the
way — hence, the top-down label. Once the defined scope of data is addressed,
a subsequent iteration is defined and the process repeats. Figure 9-4 shows the
basic data architecture for the top-down data architecture approach.
Source
Systems
Order
Entry
Product
Shipments
Financials
Capture/Create
Extract
Staging Data Store
ETL Processing*
Prepare
Data
Warehouse
Limited User
Access
Publish
ETL Processing*
Data Mart
Data Mart
Data Mart
Data
Mart
Dashboards
Use
Flat
File
Business
Intelligence
Applications
Scorecards
Reports
*Note: ETL Processes may include ETL Data Stores
Figure 9-4
Top-down data architecture approach
Capture/Create the Data
Data is extracted from the underlying source systems. These may include major
operational application systems, departmental databases, or data purchased
from third-party sources. The data architecture of this layer is determined by
the source systems.
The source system data architecture is the same regardless of the data warehouse
data architecture approach
.
Extract the Data: The Staging Data Store
Extracting data from the source systems is simply the first step in the overall
ETL system. The goal here is to acquire the data from across the organization
