Each local database can have a single Replicator Server that services all targets. If the replication traffic is heavy, performance can be affected.

Each full peer database can have a Replicator Server for each CDDS. This allows each CDDS to have a different priority (at the operating system level) and different scheduling. Also, if there is a problem with one server, the whole system is not necessarily affected.

Each target that the full peer database transmits can have its own server. This further localizes transmission problems that can occur and reduces traffic on individual servers. Many servers can improve performance over the network, however, each server uses resources. For an example of using a different server for every target database, see (see page Example: R.E.P. CDDS 0.

Even if the same target database exists in more than one CDDS, it has a different server transmitting to it for every CDDS it participates in. For an example of using different servers for every target database and CDDS, see Example: R.E.P. Server Assignments.