Glossary · System coordination, integration and orchestration
Data replication
Also known as: replication
German: Datenreplikation
In information systems, data replication is the copying and ongoing updating of data from one storage location or system to one or more others, to increase availability, bring data closer to its consumers or protect against the loss of a single copy.
- System integration
In one sentence
Data replication keeps copies of data on several systems up to date to improve availability, locality or protection against data loss.
Example
The recipe database of the site server is replicated to each line server, so lines can keep producing when the site network is down.
How it applies
- Architecture: Choose synchronous replication when no confirmed write may be lost, and asynchronous replication when latency and availability matter more. The choice determines the Consistency model readers experience.
- Operation: Monitor replication lag and failures. A replica that stopped updating continues to deliver data that looks valid but is outdated.
- Documentation: State which copy is the Source of truth, where replicas are and how long a replica may lag. Replication is not a backup: a deleted or corrupted record is replicated as well, so document backup separately.
Data replication vs. data synchronization
Replication usually copies from a primary to replicas in one direction. Data synchronization often means reconciling data between systems that can both change it, which requires conflict handling.