Hinted Handoff

A technique that improves write availability during temporary node failures — by having a healthy node temporarily accept writes on behalf of an unavailable one.

Key Takeaways

1. Hinted handoff temporarily stores writes for unavailable replicas — a nearby node holds the data and forwards it when the target recovers
2. Improves write availability during temporary node failures — writes succeed even if the designated replica is down
3. Hints are stored with a TTL — if the target node doesn't recover within the window, the hint is discarded
4. Hinted handoff is not a replacement for anti-entropy repair — it handles temporary failures, not permanent data divergence