Databases Reference
In-Depth Information
FIGURE 6.5
Datamart BUS architecture.
warehouse. After loading the data warehouse, depending on the needs of the business intelligence and
analytical applications, there are several virtual layers built with views and aggregate tables that can
be accessed by the applications. In most cases, a separate datamart is deployed to create further data
transformations.
The datamart BUS architecture shown in
Figure 6.5
builds a data warehouse from the bottom up.
In this technique, we build multiple datamarts oriented per subject and join them together using the
common BUS. The BUS is the most common data elements that are shared across the datamarts. For
example, customer data is used in sales and call center areas; there will be two distinct datamarts—
one for sales and one for call center. Using the BUS architecture we can create a data warehouse by
consolidating the datamarts in a virtual layer. In this architecture the datamarts are based on a dimen-
sional model of dimensions and facts.
Pros and cons of information factory approach
●
Pros:
●
Provides an enterprise view of the data.
●
Centralized architecture.
●
Central rules and control.
●
Refresh of data happens at one location.
●
Extremely high performance.
●
Can build in multiple steps.
●
Cons:
●
High risk of failure.
●
Data quality can stall processing data to the data warehouse.
●
Expensive to maintain.
●
Needs more scalable infrastructure.
Pros and cons of datamart BUS architecture approach
●
Pros:
●
Faster implementation of multiple manageable modules.
●
Simple design at the datamart level.
