Enterprise storage systems are specialised servers packed mostly with disks, whose software keeps extra copies of the data so that nothing is lost when a disk fails. One such system is also called a storage array. Banks, hospitals, factories and government offices keep their most important data on one.
Why not keep data on each server's own disks? If a server fails, the data on its own disks cannot be reached until it is repaired. A storage system is shared. Many servers read and write data on it over the network, so when one server fails, another server can take over its work and reach the same data. A storage system also has two controllers, the computers inside it that handle every read and every write. If one controller fails, the other takes over.
The protection against failed disks is called RAID, or, in newer systems, erasure coding. The software spreads each piece of data across many disks and adds extra pieces of calculated data, called parity. If a disk fails, the software uses the parity to rebuild the missing data onto a spare disk. This is redundancy applied to data.
Consider a chain of diagnostic labs in Nagpur that keeps every patient's X-ray and CT scan for years. Its storage system holds 24 disks of 16 TB each. The software is set up so that any two disks can fail at the same time without any data being lost. One Tuesday, disk 17 fails. The system raises an alert and starts rebuilding disk 17's data onto a spare disk, using the parity on the other disks. The radiologists keep opening scans while it works. A day later, the rebuild is complete, and an engineer replaces the failed disk.
Storage systems also take snapshots. A snapshot is a record of the data as it was at one moment, for example every hour. If someone deletes files by mistake, or if ransomware encrypts them, the data can be restored from a snapshot. Snapshots that nobody can change or delete, not even an administrator, are an important part of ransomware recovery. But a snapshot is kept on the same system. So companies still copy their data to another place with backup and recovery software.
Storage systems are built with hard disk drives, which cost less per terabyte, or with flash memory (SSDs), which are much faster. Many arrays now use only flash. Some companies skip the specialised box and run software-defined storage on ordinary servers instead. The page on data storage and backup shows where storage systems fit beside tape, cloud storage and backup.