Many storage administrators today are grappling with the question of when to use RAID 10 protection and when to use
RAID 6 uses less storage, if you have enough disks in the array.
A RAID 10 array can only store half the total disk capacity in data. The other half of the capacity is taken up by the mirror. If a RAID 6 array contains the minimum number of disks (four), then it can only hold half the total disk capacity in data, as well. The difference comes as you add disks. A RAID 10 array still dedicates half its capacity to protection. But the percentage of usable capacity increases as you add disks to a RAID 6 array. If you double the number of disks from four to eight, parity only consumes 25% of the disk capacity and it keeps going down as you add more disks.
RAID 6 is more compute-intensive.
Because RAID 6 requires two parity calculations for each write operation, RAID 6 is slower on writes than most other RAID levels. This is one reason that RAID 6 hasn't been used much. RAID controller coprocessors are often employed to handle the parity calculations and improve RAID 6 write speed.
RAID 6 can protect against two simultaneous disk failures.
Because RAID 6 uses a double parity scheme, it can protect against the failure of two disks. RAID 10 may or may not be able to protect against two disk failures, depending on where they occur. If both failed disks are in the same mirror, then the other mirror can take over. If disks in both mirrors fail, you have a problem.
RAID 10 is faster to rebuild.
The major weakness of RAID 6 is that it takes a long time to rebuild the array after a disk failure. With even a moderate-sized array, rebuild times can stretch to 24 hours, depending on how many disks are in the array and the capacity of the disks. Since RAID 6 users tend to use the biggest disks they can afford, this is an increasingly serious limitation for RAID 6.
RAID 10 doesn't require special hardware.
Most controllers support RAID 10 with good performance. If you are going to use RAID 6, it is important to use a controller specifically designed to support it. RAID 6 essentially doubles the parity calculations for every write.
RAID 10 eats up 50% of the drive capacity in mirroring the data.
The capacity penalty for RAID 6 is much smaller if you use eight or more drives in an array.
Which RAID array you choose often depends on what kind of storage you're protecting. RAID 6 is mostly found in installations using SATA drives, especially large-capacity SATA drives in their arrays. There's no technical reason why you can't use RAID 6 with SCSI drives, but there's not much incentive to do so. SCSI drives are generally smaller than SATA drives and they are considered to be more reliable.
About the author: Rick Cook specializes in writing about issues related to storage and storage management.
This was first published in January 2008