In an entry-balanced deployment some data is replicated everywhere, such as the schema, the server registry, and other shared data, and some data is replicated only on certain servers.
A replication domain contains all of the servers in a replicated topology and shares a schema. The replication domain is associated with the base distinguished name (DN) and must be a base DN of a backend.
By default, replication propagates updates to all replication servers in the topology.
However, updates to data under the entry-balancing point must be replicated only among
server instances in the same data set. Replication requires that in these deployments the
PingDirectory server is configured with a
replication-set-name global configuration property and two backends.
One backend has a base DN that is replicated globally, such as
dc=example,dc=com, and the second backend has a base DN associated
with the entry-balancing point, such as ou=people,dc=example,dc=com.
The following illustration demonstrates the global and restricted backends in a directory information tree (DIT).
If a data set name is not defined when you set up the PingDirectoryProxy server, one is provided by default. The proper configuration of an entry-balancing environment requires coordination between the PingDirectory server and the PingDirectoryProxy server. After you enable replication, you can designate the replication domain as the domain participating in entry balancing.
For more details about replication, managing the replication topology, and working with multiple backends, see the PingDirectory Server Administration Guide.