On the other side of the socket and running on the recovery server, the ABA is collecting log information. The ABA receives the redo log information in the time order it was created on the production server, and then stores this information in recovery logs. Remember, these are block storage devices that do not interact with resident file systems. As the ABA receives the data, it dynamically creates optimized State Map Transactions (SMTX). The blocks identified are then sorted in ascending device/block order. Block ordering is a more efficient organization for applying modifications to the data, or replica, on the recovery server.
Before the modifications are applied to the replica, yet another block storage device is written to with information that would allow the replica to step backward in time. This storage device is called the undo log and appears to be nothing more than a logical volume to the volume manager. Once the undo log information is saved on disk, the redo log can be applied to the replica to bring it up to date with the data on the production server.