RAID (or Redundant Array of Independent Disks) is referred to as a data storage scheme that distributes data among multiple hard drives. RAID is further divided into various levels, where each level serves different requirements for each group of users. Among all RAID levels, RAID 5 is considered as the most appropriate storage layout.
RAID 5 has gained popularity due to a slew of advantages it has over other RAID implementations. You can setup RAID 5 with a minimum of three disks at your disposal. It divides both data and parity information among three or more drives. It shares a great deal of functionality with RAID 4. However, there are some key differences. One major difference is that RAID 5 uses a distributed parity algorithm in place of the dedicated parity drive used in RAID 4. This ensures that the parity information is distributed across all drives in the array.
Distributing parity gives the advantage of removing the bottleneck caused by the dedicated parity drive. This means that you can expect a better write performance with RAID 5. Further, it gives the ability to handle parallelism required in a multi-transaction environment. RAID 5 implementation also ensures fault tolerance by storing the actual data separate from its parity information. The data block and its associated parity are stored on different member drives of the RAID.
The following is a list of some exclusive features of RAID level 5:
- RAID 5 can have a minimum of three hard drives with the same sizes and types. You do not risk your data even if one member drive in your RAID 5 array fails.
- RAID 5 usually provides excellent read performance when configured with larger stripe sizes. It is superior to RAID 0, as there is one additional drive in the array and the parity information is not used during normal reads. When considering sequential reads, it is generally good for smaller stripe sizes.
- Write performance is better than RAID 3 and RAID 4 because of the distributed parity algorithm. However, this does not suffice as the overhead encountered due to parity is still significant.
- When using hardware RAID 5 controllers that have built-in automatic rebuild and hot sparing features, you can expect good availability.
RAID 5 is a great option for those who do not wish to compromise performance and fault tolerance for a better storage efficiency. RAID 5 finds its use in a majority of relational database applications and enterprise resource planning systems. However, it is not suited for a write-heavy environment.
It is generally assumed that the failure of one drive will not affect other members of the RAID array. In reality, if one member drive fails, it can directly impact others as each one is mechanically dependent on the other. There are techniques (such as parity resynchronization or stale sub-disk recovery) that are employed to recover your lost data in RAID 5. In case one or more drives in the RAID have logical corruption, you might want to consider a professional RAID recovery software.