If one change fails but changes to data on the other databases are successful, the changes to the other databases should be rolled back to ensure that no changes are made to any of the databases. This avoids data integrity problems.
This means using a mechanism to apply the changes to all the databases or not at all. The available DBMS may be able to achieve this using a multi-phase commit, whereby the update is made and logged to each of the databases. The DBMS then checks to ensure that the updates have been made successfully. If so, the updates are committed to the databases simultaneously.
Maintaining data consistency through the use of procedure logic is critical to data, and therefore to business process integrity. Such logic should be defined by experienced designers who have an understanding of the operation of the target DBMS and operating system.
|
Copyright © 2013 CA.
All rights reserved.
|
|