Data Management - The CIO's Guide to Oracle Products and Solutions - page 198

should be apparent that data-collection standards are essential for
data integrit y.
4. Quality control, data integrity, and backup/recovery : Before data
can become corporate or shared, the appropriate mapping stan-
dards and data-definition standards must be complied with via
quality control.
Another important procedure that must be performed regularly,
especially when a project or corporate data set reaches a significant
milestone, is data backup. Data backup protects the project's prog-
ress from any unanticipated system failures or user errors by sav-
ing it to the proper media and archiving it. To ensure real-time data
integrity, recovery contingencies must also be in place.
5. Data sharing : Sharing data is essential to most projects; multiple
users must have access to the same data to make this type of data
processing effective. Before data sharing can occur, the data, meta-
data, projection, and format must all be integrated. Data sharing has
direct links with data security because certain users will have differ-
ent types of access, depending on their needs or update responsibili-
ties. Connectivity is also implicit in data sharing, which is discussed
under item 7.
6. Change management and impact analysis : Impacts from data-
standard changes or technological changes (hardware/software
upgrades) must be anticipated and planned for. The possible ramifi-
cations on certain end users, organizations, and applications must be
accounted for during a project's planning.
7. Connectivity : Connectivity is how data is shared and distributed in
a networked architecture. User requirements determine the scale of
connectivity that is required; these requirements may or may not be
apparent to the user community.
8. Input and update : The actual collection, input, and update of data
may take place in several different locations and several different
steps. The technical side of data sharing has to do with the various
data types: Different types of data require different approaches to
updating. Static data is usually low maintenance, whereas dynamic
data is usually high maintenance (monitored by the specialist, data
steward, and the database administrator). Update cycles are largely
dependent on the dynamic/static nature of the data. Input and
update protocols must be developed so that data stewards have stan-
dard methodologies to follow.