Replication is a data synchronization mechanism that ensures that updates made to a database are automatically replayed to other servers. Replication improves data availability when unforeseen or planned outages occur, and improves search performance by allowing client requests to be distributed across multiple servers.

By default, all Directory Servers participating in replication are writable, so that LDAP clients can perform updates at any of these Directory Server instances. These updates will be propagated to the other servers automatically in the background and applied in the same order as the updates were entered. The replication process flow is designed to immediately propagate changes to the other replicas in the topology with little or no latency.

The following figure demonstrates the basic flow of replication.

Basic Replication Flow
The benefits of replication can be summarized as follows:
  • High-Availability. Because the data is fully replicated on all other servers in the topology, replication allows participating servers to process all types of client requests. This mitigates any availability issues when a particular server is down due to a planned maintenance or unplanned outage. For those servers that are temporarily unavailable, they will receive updates when they become available again.
  • Improved Search Performance. Search requests may be directed to any Directory Server participating in replication, which improves search performance over systems that only access single servers. Note, however, that replication does not improve write throughput since updates need to be applied at all servers.
  • WAN Friendly Data Synchronization. The built-in compression feature in the replication protocol allows efficient propagation of updates over WAN links.