For a financial trading platform requiring zero data loss and immediate failover between two sites with latency under 2 ms, which solution fits?

For a financial trading platform requiring zero data loss and immediate failover between two sites with latency under 2 ms, which solution fits?

• A. SnapMirror Asynchronous
• B. MetroCluster
• C. FlexCache fan-out
• D. SnapMirror Synchronous

Answer: (D)

The essential requirement is zero data loss with instant handover between two sites, and the latency between sites must stay under 2 ms. The mechanism that guarantees no data loss is synchronous replication: a write is considered complete only after it has been securely written to the destination as well as the primary. This means both sites stay in lockstep, so there’s no RPO (zero data loss) and a failover can be performed with an up-to-date copy ready to take over, which is exactly what a financial trading platform needs when latency budgets are so tight. SnapMirror Synchronous provides this exact behavior, aligning the write acknowledgment with the destination and allowing immediate failover within the required latency window. The other options don’t fit as cleanly. An asynchronous replication path risks some data loss on failover because acknowledgments return before the destination is fully updated. FlexCache fan-out is about caching and accelerating reads, not maintaining a durable, synchronized copy across two sites. MetroCluster is also a robust two-site DR/HA solution, but the scenario here emphasizes the need for a precise, sub-2 ms synchronous replication path to guarantee zero data loss, which is provided by SnapMirror Synchronous.

The essential requirement is zero data loss with instant handover between two sites, and the latency between sites must stay under 2 ms. The mechanism that guarantees no data loss is synchronous replication: a write is considered complete only after it has been securely written to the destination as well as the primary. This means both sites stay in lockstep, so there’s no RPO (zero data loss) and a failover can be performed with an up-to-date copy ready to take over, which is exactly what a financial trading platform needs when latency budgets are so tight. SnapMirror Synchronous provides this exact behavior, aligning the write acknowledgment with the destination and allowing immediate failover within the required latency window.

The other options don’t fit as cleanly. An asynchronous replication path risks some data loss on failover because acknowledgments return before the destination is fully updated. FlexCache fan-out is about caching and accelerating reads, not maintaining a durable, synchronized copy across two sites. MetroCluster is also a robust two-site DR/HA solution, but the scenario here emphasizes the need for a precise, sub-2 ms synchronous replication path to guarantee zero data loss, which is provided by SnapMirror Synchronous.