While the mechanics of data backup and replication is highly dependent on the type of database, strategies boil down to two basic options: synchronized real-time replication and asynchronous replication. Asynchronous is itself a continuum that varies from being very close to synchronous to replication hours out of step. In an ideal world, we would like to have a completely up-to-date copy of our data sitting on another server so that there would be no data loss in the event of a database failure for any reason. Apart from cost, which is always a significant factor, we come up against distance, latency, and network speed as the limiting factors. Synchronous replication is practical within a data center but doesn't help if the whole data center goes down. Synchronous replication isn't practical between geographic regions but does afford the greatest protection to localized outages. When replicating to other regions, you must take into account data sovereignty implications. Depending on what legal obligations you have, replicating to another region might not be possible, or the allowed region is geographically remote. One scenario is to configure a highly available synchronous solution between data centers in an availability zone, with asynchronous replication to another Azure region. The synchronized mirror or mirrors act as read-only databases where read requests are routed to reduce the load on the primary database. The asynchronous data in the other region is the fallback position in the event of a primary region failure.