Reference Work Entry

Encyclopedia of Database Systems

pp 2397-2402

Replication for High Availability

  • Bettina KemmeAffiliated withMcGill University


Fault-tolerance; Backup mechanisms


Replication is a common mechanism to increase the availability of a data service. The idea is to have several copies of the database, each of them installed on a different site (machine). Using replication, the data remains available as long as one site is running and accessible. Fault-tolerance is related to availability. A system is considered fault-tolerant if it continues to work correctly despite the failure of individual components. Replicating data and processes over several sites, the failure of any individual site can be masked since the tasks executed by the failed site can be transferred to one of the available sites. The terms high availability and fault-tolerance are often used interchangeably. However, fault-tolerance is stronger than a high availability solution since it expects the fault-tolerant system to behave exactly as a system where components never fail. This requires to make failures transpare ...

This is an excerpt from the content