A company requires an RPO of less than 10 seconds to ensure business continuity. Which technology should be deployed?

A single data center with redundant hardware protects against component failure but cannot meet a sub-10-second RPO in the event of a facility-wide outage such as a power failure or fire.

Asynchronous replication introduces a lag between the primary and secondary site that may range from seconds to minutes depending on I/O load and network latency, making a guaranteed sub-10-second RPO impossible.

Synchronous replication requires every write to be committed at both the primary and secondary site before the transaction is acknowledged, resulting in an RPO that is effectively zero or measured in milliseconds - well under 10 seconds. Geographic dispersion protects against site-level disasters that would otherwise cause complete data loss. This is the only listed option that can guarantee sub-10-second RPO regardless of failure scenario.

A single data center with duplicated infrastructure and dual PSUs provides hardware-level redundancy only and has no off-site data copy, meaning a site-level disaster would result in data loss far exceeding 10 seconds.