Before data replication can start, the target database must have a copy of the database objects that are going to be replicated (this can also be referred to as seeding the target database). It is also highly desirable that the data in the objects to be replicated are also completely in sync. If the data is not in sync before replication starts, there is a strong possibility that data conflicts can occur. For instance if a record is updated or deleted on the source database that does not exist on the target database.
