A SQL Server replication type that allows data changes to be made independently at both the publisher and subscriber, with conflicts detected and resolved when changes are synchronized.
- Merge replication is designed for occasionally connected or disconnected environments: It is well suited for mobile workers, remote offices, and other scenarios where bidirectional updates are required.
- Conflict resolution in merge replication is handled by configurable conflict resolvers: The default resolver gives priority to publisher changes, although custom conflict resolvers can also be implemented.
- Unlike transactional replication, which propagates changes only from publisher to subscriber, merge replication allows both sides to make changes independently: These changes are synchronized when connectivity is available.
- Merge replication adds tracking columns and triggers to published tables: DBAs must account for the additional storage and performance overhead introduced by these objects.
- Common use cases for merge replication include: Field service applications, retail point-of-sale synchronization, and multi-site data consolidation where network connectivity may be intermittent.
- Related terms: Transactional Replication, Replication Monitor, Snapshot Replication, Publisher, Subscriber.
