Databases Reference
In-Depth Information
Replication and Data Transport
The previous section discussed the use of multiple database servers acting together as
part of a single logical database for users. In this section, we cover data replication and
data transport in order to duplicate data.
Replication techniques are frequently used:
• When data available locally eliminates network bandwidth issues or contention for
system resources
• When mobile database users can take their databases with them and operate dis‐
connected from the network
• When redundant databases can help to deliver higher levels of reliability, as each
database can be used as a backup for other databases
• During a database or application migration
In grid implementations, the ability to share resources across the grid of computer
systems and databases can also require data to be replicated to multiple servers within
the grid.
The biggest issue facing users of multiple identical or similar databases is how to keep
the data on all of the servers in sync as the data is changed over time. As a user inserts,
updates, or deletes data on one database, you need to have some way to get this new
data to the other databases. In addition, you will have to deal with the possible data-
integrity issues that can crop up if the changes introduced by distributed users contend
with each other.
Oracle offered a number of strategies to address this situation over the years, including
Advanced Replication in the Oracle Database, Oracle Advanced Queuing (AQ), and
Streams. In Oracle Database 12
c
, some of Oracle's features still leverage AQ and Streams
under the covers. However, a number of years ago, Oracle began guiding its customers
toward using Oracle GoldenGate for asynchronous replication and Data Guard for
synchronous replication.
Replication Basics
The copying and maintaining of database tables among multiple Oracle Databases on
distributed systems is known as
replication
. Changes that are applied at any local site
are propagated automatically to all of the remote sites. These changes can include up‐
dates to data or changes to the database schema. Replication is frequently implemented
to provide faster access for local users at remote sites or to provide a disaster-recovery
site in the event of loss of a primary site.
