Database Reference
In-Depth Information
Most of the benefits of the data warehouse will not be realized in the first
delivery. The first project will be the foundation for the next, which will in
turn form the foundation for the next. Data warehousing at the enterprise
level is a long-term strategy, not a short-term fix. Its cost and value should
be evaluated across a time span sufficient to provide a realistic picture of
its cost-to-value ratio.
The following seven components make up the enterprise data warehouse
architecture. These components offer a high level of flexibility and scalabil-
ity for the enterprise wishing to implement a business intelligence solution.
Source Systems
A data-source system is the operational or legacy system of record whose
function it is to capture the transactions of the business. Source systems
should be thought of as outside the data warehouse, because we have no
control over the content and format of the data. The data in these systems
can be in many formats, from flat files to hierarchical and RDBMS, etc.
Other sources of data may already be cleansed and integrated and avail-
able from operational data stores.
Data Staging Area
The data staging area is the portion of the data warehouse restricted to
extracting, cleaning, matching, and loading data from multiple legacy sys-
tems. The data staging area is the back room and is explicitly off limits to
the end users. The data staging area does not support query or presenta-
tion services. A data-cleansing tool may be used to process data in the
staging area to resolve name and address misspellings and the like, as well
as resolve other data-cleansing issues by use of fuzzy logic.
Data Warehouse Database
The warehouse is no special technology in itself. The data warehouse data-
base is a relational data structure that is optimized for distribution. It col-
lects and stores integrated sets of historical, nonvolatile data from multiple
operational systems and feeds them to one or more data marts. It becomes
the one source of the truth for all shared data.
Search WWH ::
Custom Search