databasereplication

Database replication is the process of copying and maintaining database objects, such as data and logs, across multiple database instances to improve availability, fault tolerance, and read scalability. Replication can also support disaster recovery and geographic distribution of data.

Common architectures include master-slave (primary and one or more replicas), master-master or multi-master (multiple writable nodes), and cascading or peer-to-peer configurations. In master-slave setups, writes occur on the primary and are replicated to the replicas. In multi-master configurations, any node may accept writes, requiring conflict detection and resolution to maintain consistency.

Replication modes describe how closely replicas follow the primary. Synchronous replication waits for acknowledgments from replicas before completing a transaction, providing strong consistency at the cost of higher latency. Asynchronous replication returns results to the client before replicas acknowledge, reducing latency but allowing potential lag and temporary inconsistency.

Replication mechanisms determine how changes are captured and delivered. Common methods include log-based replication (using write-ahead or binary/transaction logs), statement-based and row-based replication, streaming replication, and change data capture. Replication can be configured as push (publisher to subscribers) or pull (replicas pull changes).

Data scope and topology can be full or partial, with options to filter tables or rows. Schema changes may be replicated automatically in some systems but often require separate handling. In multi-master environments, conflict resolution strategies—such as last-writer-wins, version vectors, or application-defined policies—are essential to maintain data integrity.

Use cases span read scaling, high availability, disaster recovery, and geographic distribution. Considerations include latency, bandwidth, failure scenarios, data sovereignty, security, and monitoring overhead. Common examples include PostgreSQL streaming replication, MySQL replication, and SQL Server or Oracle high-availability solutions.