Redundant data is another option for dealing with primary site failure. Implementing a redundant data approach differs from using a standby database, which duplicates the entire primary database. Data redundancy is achieved by having a copy of your critical data in an entirely separate Oracle database with a different structure. The data, not the database itself, is redundant. If the primary site fails, users can continue working using the redundant data in the secondary database.
Oracle provides automated synchronous and asynchronous data-replication features to support data redundancy. For simplicity, in the following sections we’ll examine replication using a simple two-site example—a primary and a secondary. Oracle can, however, perform N-way or multimaster replication involving more than two sites with all sites replicating to all others.
Whenever you have a data replication scenario, you always have a primary site, from which the replication originates, and a secondary site, which is the recipient of the data replication. (In a multimaster scenario, you can have more than one master site, and a single machine can be a master for one replication plan and a secondary site for another.) When you design your replication plan, you must consider the degree to which data at the secondary site can differ for a period of time from the data at the primary site. This difference is referred to as data divergence