Why scale-out databases still need backup despite replication

Human and application errors are part of this. Accidental deletions, faulty updates and misconfigurations are common, and replication cannot tell an intentional change from a mistake. Once the mistake happens, it propagates everywhere.

Replication also gives you no point-in-time recovery. It keeps systems synchronized in the present and cannot take you back to a known good state from yesterday or last week. Backups and snapshots provide that historical recovery.

Security is another concern. In a ransomware attack or other malicious activity, replication will often copy the encrypted or compromised data to every node. Without isolated backups, you have very few ways to recover.

Compliance and audits add their own requirement. Many industries require long-term data retention and the ability to recover historical states. Replication is about keeping the current state consistent, so it does not preserve history.

The history of infrastructure makes the same point. RAID, clustering and replication each improved availability over time, and none of them removed the need for backup. Scale-out databases follow the same pattern.

Snapshots and backups add what replication lacks: point-in-time recovery, isolation from failures at the source, and protection against logical errors. Replication keeps systems running and backup lets you recover when something goes wrong, so you need both.