Enterprise Data Management - High-Performance Web Databases: Design, Development, and Deployment - page 31

Exhibit 3-5. Optimum sourcing processes.

assess effects, and direct the appropriate changes in the enterprise ware-
house. The business data owners are responsible for this process, for resolv-
ing any conflicts that result from data definitions and requirements, and for
ensuring that the implementation of EDM meets their business needs.

The optimum source data and its operational process must be managed
because by the nature of the extracting copy management processes, the
operational data can affect propagation of the enterprise warehouse. It is
this process that feeds the enterprise warehouse in an assortment of ways.
These transport activities can occur at any point within the lifecycle of
delivering data from the optimum source to any location managed within
the enterprise data concept.

The following processes, shown in Exhibit 5, are not limited to optimum
sourcing and can recur any time the data is extracted or distributed within
the EDM process:

• Refresh versus net change. The refresh process entails all the data with-

in a database or file being replaced on a periodic basis; with net change
only those elements that had changes are affected (i.e., how much).

• Drip versus batch. In the drip process, data movement occurs continu-

ously on a recurring basis as warranted by modifications, in batch
mode, data is bulk-processed at end of completed operational pro-
cessing (i.e., when).

• Pull versus push. Data can be pulled from operational sources as initi-

ated by EDM, or data elements or fields can be pushed down from the
operational systems when changes are noted (i.e., who initiates).