The solution combines at least two drives to create a storage pool. JBOD storage pools do not offer data redundancy. The available capacity of a JBOD storage pool equals the total capacity of all drives included in the storage pool. JBOD supports combining drives of different sizes.
RAID 0 combines two or more drives to increase performance and capacity, but provides no fault tolerance. A single drive failure will result in the loss of all data on the array. RAID 0 is useful for non-critical systems where a high price/performance balance is required.
RAID 1 is most often implemented with two drives. Data on the drives is mirrored, providing fault tolerance in case of drive failure. Read performance is increased while write performance will be similar to a single drive. A single drive failure can be sustained without data loss. RAID 1 is often used when fault tolerance is key, while space and performance are not critical requirements.
RAID 5 provides fault tolerance and increased read performance. At least three drives are required. RAID 5 can sustain the loss of a single drive. In the event of a drive failure, data from the failed drive is reconstructed from parity striped across the remaining drives. As a result, both read and write performance are severely affected while a RAID 5 array is in a degraded state. RAID 5 is ideal when space and cost are more important than performance.
RAID 6 is similar to RAID 5, except it provides another layer of striping and can sustain two drive failure. A minimum of four drives is required. The performance of RAID 6 is lower than that of RAID 5 due to this additional fault tolerance. RAID 6 becomes attractive when space and cost are important and sustaining multiple drive failures is required.
RAID 10 combines the benefits of RAID 1 and RAID 0. Read and write performance is increased, but only half of the total space is available for data storage. Four or more drives are required making the cost relatively high, but the performance is great while providing fault tolerance at the same time. In fact, a RAID 10 can sustain multiple drive failures—provided the failures are not within the same subgroup. RAID 10 is ideal for applications with a high input/output demand such as database servers.
RAID 50, also known as RAID 5+0, merges distributed parity (RAID 5) with striping (RAID 0) and requires a minimum of six drives. The benefits of this RAID level are better write performance, better data protection, and faster rebuilds than RAID 5. Performance does not degrade as much as in a RAID 5 array because a single failure only affects one array. Up to four drive failures can be overcome if each failed drive occurs in a different RAID 5 array.
RAID 60 (or RAID 6+0) is a hybrid that offers the distributed double parity of RAID 6 with the straight block-level striping of RAID 0. As a RAID 0 array striped across RAID 6 elements, RAID 60 requires eight drives at minimum. RAID 60 (6+0) is a multilevel disk set, composed of RAID 6 sets aggregated at a higher level into a RAID 0 array. A RAID set offers redundancy and can withstand the loss of up to two disks in each parity set. RAID 60 arrays are more reliable than RAID 50 arrays thanks to the extra parity disk in RAID 60. But when more than two disks in a single parity set are lost, the RAID 0 set breaks, and data recovery is needed.