What are the pros and cons of RAID-6, and how does this compare to proprietary forms of RAID-6 (dubbed RAID-DP)? |
 |
| 08 May 2007 | Greg Schulz |
|
|
Generally speaking, dual-parity RAID (RAID-DP) is a proprietary, vendor specific implementation of dual parity. RAID-6 dual parity is nothing new; it's been around since the 1990s. It never really took off because of the added drive cost and performance impact. However, RAID-6 and its proprietary versions are finally being embraced because of the high rebuild times of today's high-capacity disk drives. Dual parity protects a RAID set against the possibility of a second disk failure before the first disk rebuilds. It's important to note that while RAID-6 is often discussed in conjunction with SATA drives, RAID-6 applies to all drive types, including SAS and Fibre Channel.
So, what is different today? The capacity of today's disk drives is much higher, while their cost is far less. The processing power of today's RAID controllers is also much greater, so you can actually have RAID-6 parity calculations offloaded from the main processor, so there is no performance impact on the server or array as their might have been years ago. Still, there is a performance penalty for the additional parity and the cost of the additional drive to retain the second parity data -- along with the all the other overhead, such as the drive tray space, power and other support needed within the array itself.
When you're considering RAID-6, examine how it performs with different workloads (e.g. sequential vs. random), and determine how it is optimized for rebuilds.
Go back to the beginning of the Disk Hardware FAQ Guide.
|
|
|
|
