An administrator must replicate a protection domain to a remote Nutanix site over a WAN link with 80 ms latency. The business tolerates a one-hour RPO and cannot accept a synchronous commit path. Which replication mode best fits this DR design?

Asynchronous replication takes scheduled snapshots and ships them to the remote site in the background, so VM writes are acknowledged locally and never wait on the WAN. Its schedules start at one hour, which matches the stated RPO, and it tolerates high-latency links because replication time only affects how fresh the remote copy is, not application performance. NearSync is also asynchronous from the application's point of view: it ships lightweight snapshots every 1 to 15 minutes and does not require a synchronous commit path or very low latency. It is simply aimed at a much tighter RPO and needs enough bandwidth to keep up with the change rate, so it adds cost and load that a one-hour RPO doesn't justify. Metro Availability writes synchronously to both sites, which requires round-trip latency of 5 ms or less; at 80 ms every write would stall, and the stem forbids a synchronous path. A witness adds automated failover decisions to Metro, but the data path is still synchronous, so it does not remove the latency requirement. Exam caveat: map the RPO first: zero means Metro, 1 to 15 minutes means NearSync, and an hour or more means async. Operational check: confirm the measured RTT and bandwidth, set an hourly schedule, and verify the age of the latest recovery point at the remote site.