A financial services company is designing a disaster recovery solution for its core transactional database. The business has a zero-tolerance policy for data loss, meaning the Recovery Point Objective (RPO) must be zero. The primary data center and the disaster recovery site have a high-bandwidth, low-latency dedicated fiber connection. Which replication method BEST meets this requirement?
The correct answer is synchronous replication. Synchronous replication writes data to both the primary and secondary storage locations simultaneously. A write operation is not considered complete until an acknowledgment is received from both sites. This method guarantees that the data at the disaster recovery site is an exact mirror of the primary site at all times, achieving a Recovery Point Objective (RPO) of zero. Given the strict requirement for no data loss, this is the most appropriate choice.
Asynchronous replication is incorrect because it writes data to the primary storage first and then copies it to the secondary site later. This introduces a delay, meaning that some data could be lost if a disaster occurred before the replication was complete, resulting in an RPO greater than zero.
Snapshot replication is a form of asynchronous replication where point-in-time copies are made and sent to the DR site periodically. This would not meet the zero RPO requirement, as any data changed between snapshots would be lost.
Bidirectional replication allows for data to be written and replicated in both directions between two active sites. While useful in active-active scenarios, it does not specifically address the zero data loss requirement for a primary-to-DR-site setup as directly as synchronous replication does.