RAID 0 - Provides higher performance and additional storage, but does not include any redundancy. If any drive in the array fails the array is destroyed. This risk increases with the addition of drives to the array.
RAID 1 - Commonly referred to as striping; data is identically written or mirrored to two drives. The read requests are service by the drive with the least seek time plus latency from rotation. The write performance will be determined by the slower drive. RAID 0 will continue to operate as long as at least one drive is functioning.
RAID 5 - Requiring at least three disks to operate parity and data is distrubuted and the array is not destroyed if a single drive is lost. Reads following a drive loss are calculated from the parity causing the drive to be masked from the end user. The rebuilding of an array after a failure will add stress to all of the working drives as all areas of the disk considered in use will need to be read to rebuild the redundancy. This can result in other drives that were close to failure to fail before the rebuild is finished.
RAID 6 - Providing fault tolerance of up to two drives RAID 6 is more practical for larger groups of drives. The larger the drive capacity, the longer amount of time it will take to recover from a drive failure, which will also reduce performance of the entire array until the failed drive is replaced and the array rebuilt.
RAID 10 - Is somewhat of a combination of RAID 0 and RAID 1 with all data being written across the primary disk in a striping fashion and mirror on the secondary disks.
There are also other types of RAID (2,3, & 4), but they are either theoretical or rarely used or not considered a "standard level."
No comments:
Post a Comment