Replication is a method of data backup in which the same data is copied from a primary storage medium to one or more secondary storage media.
The replication procedure is also used for data reconciliation in distributed databases and ensures that all database tables are identical. The distributed databases are reconciled with the original database on a regular basis or online. This ensures data consistency between the distributed databases. The replication method is used for transactions, which means that the databases have the same data sets.
Replication is divided into synchronous, asynchronous and bidirectional replication. Insynchronous repl ication, changes in the data of the primary storage are copied simultaneously to all secondary stor ages. As a result, all stores, primary and secondary, have current and identical data sets at all times. Synchronous data exchange can take place via storagearea networks ( SAN), local area networks( LAN) or wide area networks(WAN). The synchronous method is limited to distances of a few hundred kilometers because the run times result in high latencies.
The situation is different with asynchronous replication. With this technique, the data in the primary storage is updated and stored first, and only after a certain latency period are the data files in the backups updated. The files can be transferred using the FTP protocol or the SSH protocol.
Bidirectional repl ication is about replication of data between primary storage and backup, where the functions of the storage media are interchangeable and the data can be exchanged in both directions between the storage media. In bidirectional replication, the primary storage and backup storage have interchangeable roles. In contrast, in unidirectional replication, the roles of the storage media are fixed. One storage medium is the primary storage, another is the backup storage. Data can only be transferred from the primary storage to the backup storage.