A company has an RPO of 30 minutes for a 20 TB file-server share replicated to a secondary data center 600 km away over a high-latency WAN link. Real-time mirroring has already caused noticeable application slowdowns because each write must traverse the 120 ms round-trip link. To meet business requirements, the administrator configures the storage array to acknowledge writes locally and then copy the changed blocks to the DR array every 15 minutes without pausing production I/O. Which replication approach best matches this design?
Bidirectional replication that permits write operations at either site and reconciles conflicts
Constant replication that continuously streams every write to the DR site in real time
Background (asynchronous) replication that batches changed blocks after local commits
Synchronous replication that requires remote acknowledgment before completing each write
The scenario describes an asynchronous process in which the production array first commits the data locally, immediately answers the application, and then transfers the accumulated changes on a schedule. CompTIA's disaster-recovery objective calls this background replication. Because it runs after the write completes, it avoids the latency penalty of waiting for remote acknowledgment, but some data (up to 15 minutes in this case) could be lost if the primary site fails before the next cycle.
Constant replication would stream every write to the target as soon as it occurs, increasing bandwidth use and again tying application commit time to WAN latency. Synchronous replication explicitly waits for the remote array to confirm each write, causing the performance issue the administrator is trying to avoid. Bidirectional mirroring allows writes at both sites and adds conflict-resolution complexity, which is not part of the stated requirement.