Research and official database documentation.

PostgreSQL

• PostgreSQL: High Availability, Load Balancing and ReplicationOfficial overview of standby, replication and load-balancing approaches.
• PostgreSQL: Logical ReplicationOfficial documentation for publication/subscription-based logical change replication.
• PostgreSQL: Logical Replication ConflictsConflict types, error behavior and operational handling.
• PostgreSQL: Comparison of Replication SolutionsOfficial comparison of physical, logical and multi-directional approaches.

MySQL

• MySQL InnoDB Cluster IntroductionOfficial overview of InnoDB Cluster, Group Replication and single- or multi-primary operation.
• MySQL Group ReplicationOfficial replication architecture and group membership documentation.
• MySQL Group Replication RequirementsIncludes primary-key requirements and engine constraints.
• MySQL Group Replication LimitationsVersion-specific limitations and multi-primary considerations.

MariaDB Galera Cluster

• MariaDB Galera Cluster GuideOfficial multi-primary cluster overview.
• MariaDB Galera ArchitectureGroup communication, write-set certification and replication architecture.
• MariaDB Galera Replication GuideOperational context for multi-primary read/write participation.

Cloud and architecture references

• AWS: Understanding Active-Active ConflictsConflict context for PostgreSQL active-active replication on Amazon RDS.
• Azure Architecture Center: Sharding PatternShard-key, distribution and operational pattern guidance.
• Oracle GoldenGate: Active-Active vs Active-PassiveOfficial comparison in a replication context.

Foundational research context

• Coordination Avoidance in Database SystemsResearch on when coordination is required to preserve application invariants.

Research results and database products operate under specific assumptions. ShardronDB public pages use these sources to frame questions, not to claim identical guarantees.

Editorial and evidence policy

• Technical claims should link to official documentation or a reproducible test.
• Version-specific behavior should name the tested database and version.
• Performance claims should include hardware, dataset, workload, topology and methodology.
• Conceptual benefits should use terms such as “may”, “can” or “deserves evaluation” until independently measured.
• Pages should state limitations and suitable alternatives.
• Reference pages should be reviewed at least quarterly or after major database releases.

Last editorial review: 2026-07-26.