Data independence

The structures at each level of the ANSI-SPARC architecture are independent of structures at other levels. Thus one level may change without having an impact on any other level - the thing that does need to change is the mapping from the modified level to a related level, and that is handled transparently by the DBMS.

For example, the way a department processes some of the data may change in such a way that requires the data to be presented differently. The data is the same, but the presentation is different; therefore, there is no change required at the conceptual level. Likewise, a DBA may reorganise the files in which the data is stored, but again does not require any changes at the conceptual level.

The ANSI-SPARC architecture therefore delivers two types of data independence as illustrated in the diagram below.

Data independence means that the extent of any change in the database structure is limited and can be assigned to someone in an appropriate role with the appropriate knowledge. Applications that do not exhibit data independence have no such clean divisions, and so a simple conceptual change - changing the length of a stored data item, for example - might require many individual updates throughout the application. Changes to the way in which data is physically stored could affect logical structures and user views. Maintenance is therefore inefficient because the entire application must be checked through even for the simplest modification.

This module is mainly - but not exclusively - concerned with the conceptual level.