You can protect your data through cross-region volume replication, which asynchronously replicates a source volume in one region to a destination volume in a different region. This capability enables you to use the other volume for critical application activity in case of a region-wide outage or disaster.
Volume replication moves only used data blocks during the initial transfer. Subsequently, only changed blocks are transferred during incremental transfers. Charges are incurred only for bytes transferred, which optimizes transfer times and lowers costs.
Volume replication workflow
During volume replication, all source volume content is replicated to the destination volume in a process called initial transfer. The initial transfer takes a snapshot on the source system and transfers its contents to the destination volume. After the initial transfer completes, the replication mirror status changes to mirrore. As a result, the destination volume becomes read-only and reflects the contents of the source volume's snapshot, including all snapshots taken before the initial snapshot.
After the initial transfer is complete, the scheduled replication interval proceeds in the form of incremental updates. First, a new snapshot is created on the source. Then, the data changed between the new and the previous snapshots is calculated. Finally, these changes are transferred to the destination volume. During this process, the transfer status appears as transferring in the user interface. After all the changes are transferred, the destination volume's contents transitions from the old snapshot to the new snapshot.
Considerations and limitations
Review the following considerations and limitations for volume replication use:
Stopping replication relationships
You can stop a replication relationship between the source and destination volume. Stopping a replication relationship changes the replication relationship status to stopped and makes the destination volume read-write and usable. Both volumes continue to share the latest replication snapshot as a new baseline for later resynchronization.
Resuming stopped relationships
You can resume a stopped relationship. When you resume a replication relationship, the destination volume reverts back to the latest initial replication snapshot, at which point, the volume replication process starts over.
When you resume a replication relationship, any changes made to the destination volume while the replication was stopped are deleted.
When you stop a replication relationship after initialization, you can reverse the direction of the replication by swapping the source and destination volume roles.
Deleting replication relationships
You can only delete stopped replication relationships with a stopped status. Deleting replication relationships removes initial replication snapshots. You can't resume a deleted relationship.
Replication pricing
Replication is charged separately from volume capacity, and is based on the number of bytes transferred between primary and secondary volumes. For more information, see Billing calculation for volume replication.