Enterprise Storage Forum content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
RAID 5 and RAID 6 are two different levels of Redundant Array of Independent Disks, a data storage technology for disk drives. RAID 5 and 6 are very similar; both use a combination of disk striping and parity to store data across an array of drives. Disk striping distributes file data across all drives in a RAID array, and parity uses mathematical equations to recalculate data if a drive should fail.
Technical Specifications for RAID 5 and RAID 6
RAID 5 arrays require a minimum of three disks. RAID 6 arrays require a minimum of four disks. Although maximum array sizes are dependent on controllers and other limitations, 32 disks is often listed as the limit for RAID 5 and RAID 6 arrays.
RAID 6 is similar to RAID 5—both use disk striping and parity. Where RAID 5 only has one instance of parity, RAID 6 has two. This allows a RAID 6 array to withstand two drive failures rather than one. The second instance of parity is much more complex than the first (which RAID 5 also uses).
Read speeds for RAID 5 and RAID 6 are similar to the speed of other RAID levels. However, if a disk has failed, read speeds will decrease to compensate for the missing data.
Write speed is where these two levels sometimes take a hit: because recalculating and rewriting a disk from parity takes time, write speeds can be rather sluggish.
Fault Tolerance in RAID 5 vs. RAID 6
RAID 5 has one instance of parity spread across the disks in an array. Because it only uses one type of parity, it only allows for one disk failure. If one disk fails, then the parity in the array can be used to rebuild the missing data.
However, if two disks fail at the same time, some of the data stored in the array of disks will be lost. Because it only allows one disk failure, RAID 5 is considered by some storage experts to no longer be a good choice for a new RAID configuration.
RAID 6 has two instances of parity. One is the common XOR operation, and the other is a more complicated method of recalculating data. RAID 6 allows for two disk failures. If two disks fail at the same time, the parity distributed across the disks can rebuild the data. But if three or more disks fail, data will be permanently lost.
Criticism of RAID 5 and RAID 6
Some storage researchers believe that neither RAID level is relevant, due to the possibility of multiple-drive failure. RAID 5’s fault tolerance only allows for one disk failure before data loss ensues, since RAID 5 only has one instance of parity. If someone has a very large RAID 5 array, say twenty disks, there is a possibility that more than one of those drives will fail. Data would then be lost.
In hard disk drives, an unrecoverable read error (URE) is a section of a drive that is completely lost. It’s a ruined sector on a disk that cannot be rebuilt. When a disk in an array fails, and the RAID 6 rebuilding process begins through parity, there is a possibility that a URE will occur on disks while data is being reconstructed.
The larger an array is, the longer it will take to rebuild data once a drive is lost. The likelihood of finding another URE during a rebuild process increases with the size of the array.
Some storage experts believe that neither RAID 5 nor RAID 6 are reliable file storage or backup choices compared to increasingly popular cloud storage. However, many individuals and enterprises still use RAID for their on-premises hard drive file storage. For those who don’t want to pay for expensive cloud storage plans and prefer to stay with legacy systems, disk arrays and RAID controllers are still an attractive file storage option.
