Replication is the server process of copying data modifications from one location to another. The main characteristics of replication which make it appealing are:
The unit of replication is a transaction, not just an individual sequel (SQL) statement.
Replicated transactions are applied in the same order as they occurred on the primary side.
The replication system is able to detect whether a network connection is temporarily offline and when the component is available again. When a disruption like this occurs, the replication process should continue functioning without having to be adjusted.
Although the replication server is powerful and flexible, it is not easy to administrate. Another contributing factor is that the scope and complexity of the database administrator's job expands significantly when replication is involved. The previously independent data servers are now closely related because data is replicated between them. This effectively means that the data servers have become one large system. Thus, making a database administrator's job much more difficult.