Learning About Data Guard - Expert Consolidation in Oracle Database 12c - page 328

Figure 9-3. Example for cascaded standby databases.
In Oracle 12.1 the database CDB1DCB cascades redo as soon as it arrives in the standby redo logs. From a
configuration point of view there are two pairs of configurations. The first pair is between CDB1DCA and CDB1DCB,
and the second pair between CDB1DCB and CDB1DCC1 plus CDB1DCC2. As a result the database-in-the middle will
be configured as the communication endpoint for CDB1DCA. Similarly the two standby databases in data center C
will be the communication endpoint for CDB1DCB.
In previous releases Oracle did not pass redo on directly. Instead you had to wait until the standby redo log of the
cascading database (CDB1DCB) had the log archived. The enhancement in Oracle 12c is named real-time-cascade
and requires the Active Data Guard license. The old way is still available although you can cascade to a maximum of
10 databases that way. This limit does not exist for real-time cascade.
The use of cascaded standby database is not anticipated in the hosting services and therefore will not be covered
in further detail here.
The Far Sync Standby Feature
The far sync standby is another approach to ease the burden of redo shipping to multiple standby databases, as
shown in Figure 9-4 . The benefit of the far sync standby is increased with every additional standby database in
the configuration. The Far Sync standby is implemented as a cascaded destination. Despite the name, the far sync
standby does not have to be far away, it could be geographically close and still help a lot with distributing redo.