Save the saves – RAID array |
|
Each new RAID scheme builds upon the scheme before it and adds in important features which improve performance and resilience. RAID 0 does not provide any additional data loss protection and is for improving read write time by splitting the work load across multiple disks. RAID 1 mirrors the primary disk across secondary disks but provides very little improvement in speed. RAID 1 and higher schemes can survive the failure of a disk in the array without loss of data, though performance suffers. RAID 2 uses bit level striping for speed and introduced error correction improvements with a space intensive parity code to detect and repair errors. RAID 3 uses byte level striping and introduced a dedicated parity disk for more efficient error correction. RAID 4 stripes data in full blocks instead of bytes, thus allowing a second block from a separate drive to be accessed simultaneously and also uses a dedicated parity disk. If the parity disk fails on a RAID 3 or 4 setup, the error checking and disk reconstruction abilities of the array are lost until a new parity disk is installed.
The current standard RAID setup is RAID 5 which uses the same block striping of RAID 4 schemes but also stripes the parity data in order to remove the bottleneck found in the dedicated parity disk of RAID 3 and 4 schemes. RAID 6 is similar to RAID 5 in almost all aspects but provides redundancy of the parity data spread across the drives in order to prevent the loss of parity information if a drive containing parity information fails. A RAID 6 setup can actually survive the loss of two disks.
A RAID array provides much more resiliency and speed to the storage scheme of a computer system. RAID is not, however, the only precaution needed to protect against data loss. A secure offsite backup of critical data is still an absolute necessity to prevent loss in the event of a natural or man-made disaster destroying or corrupting the system's primary storage. |
|