Editing Computer data storage (section)

=== Redundancy ===
{{Main|Disk mirroring|RAID}}
{{See also|Replication (computing)#Disk storage replication|label 1=Storage replication}}

While a group of bits malfunction may be resolved by error detection and correction mechanisms (see above), storage device malfunction requires different solutions. The following solutions are commonly used and valid for most storage devices:

* '''Device [[Disk mirroring|mirroring]] (replication)''' – A common solution to the problem is constantly maintaining an identical copy of device content on another device (typically of the same type). The downside is that this doubles the storage, and both devices (copies) need to be updated simultaneously with some overhead and possibly some delays. The upside is the possible concurrent reading of the same data group by two independent processes, which increases performance. When one of the replicated devices is detected to be defective, the other copy is still operational and is being utilized to generate a new copy on another device (usually available operational in a pool of stand-by devices for this purpose).
* '''Redundant array of independent disks''' ('''[[RAID]]''') – This method generalizes the device mirroring above by allowing one device in a group of devices to fail and be replaced with the content restored (Device mirroring is RAID with ''n=2''). RAID groups of ''n=5'' or ''n=6'' are common. ''n>2'' saves storage, when compared with ''n=2'', at the cost of more processing during both regular operation (with often reduced performance) and defective device replacement.

Device mirroring and typical RAID are designed to handle a single device failure in the RAID group of devices. However, if a second failure occurs before the RAID group is completely repaired from the first failure, then data can be lost. The probability of a single failure is typically small. Thus the probability of two failures in the same RAID group in time proximity is much smaller (approximately the probability squared, i.e., multiplied by itself). If a database cannot tolerate even such a smaller probability of data loss, then the RAID group itself is replicated (mirrored). In many cases such mirroring is done geographically remotely, in a different storage array, to handle recovery from disasters (see disaster recovery above).