Your company operates a revenue-critical OLTP database on a SAN in its primary datacenter and has a secondary datacenter 25 km away connected by redundant 25 Gbps dark-fiber links with round-trip latency below 2 ms. Management has set a recovery point objective of 0 seconds and a recovery time objective of 5 minutes. Administrators want the simplest solution that meets these objectives without installing any host-level agents or clustering software. Which replication technique BEST satisfies these requirements?
Host-based asynchronous file replication using rsync every 5 minutes
SQL Server log shipping to the secondary site at 5-minute intervals
Hourly snapshot replication to a public-cloud object store
Array-based synchronous block-level mirroring between the two SAN controllers
Array-based synchronous block-level mirroring writes each block to both storage systems and waits for acknowledgment from the remote array before confirming the write to the application, so no data can be lost (RPO = 0) and failover can occur almost immediately (very low RTO). Because the function is built into the SAN controllers, it adds no host-side complexity. Host-based rsync and SQL Server log shipping are asynchronous; they acknowledge writes locally and therefore allow minutes of data loss. Hourly snapshot replication can lose up to an hour of data and involves additional cloud components, so none of those options meet the stated objectives.
Ask Bash
Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.
What is synchronous block-level mirroring?
Open an interactive chat with Bash
What do RPO and RTO mean in disaster recovery?
Open an interactive chat with Bash
Why is array-based replication preferable to host-based methods like rsync?