Saturday, January 30, 2010

"Single digit" RAID Types

"RAID" is now used as an umbrella term for computer data storage schemes that can divide and replicate data among multiple hard disk drives. The different schemes/architectures are named by the word RAID followed by a number, as in RAID 0, RAID 1, RAID  5 etc. RAID's various designs involve two key design goals: increase data reliability and/or increase input/output performance. When multiple physical disks are set up to use RAID technology, they are said to be in a RAID array. This array distributes data across multiple disks, but the array is seen by the computer user and operating system as one single disk. RAID can be set up to serve several different purposes.

RAID 0 (Disk striping)

  • Drives required (minimum): 2
  • Max capacity: Number of disks x disk capacity
  • Description: Data to be written to the disk is broken down into blocks with each block written to a separate disk.
  • Pros: Very, very fast since data is written to and read from storage over multiple "spindles", meaning that the I/O load is distributed. The more disks that are added, the better the performance (in theory). As always, if you’re looking for huge performance gains, use a tool such as IOmeter to test your storage performance as the gains may not be that great.
  • Cons: When a single drive fails, the entire array can be compromised since this RAID level does not include any safeguards. As disks are added, the risk of failure increases.

RAID 1 (Disk mirroring)

  • Drives required (minimum): 2 (or multiples of 2)
  • Max capacity: Total array capacity divided by 2
  • Description: All data that is written to the storage system is replicated to two physical disks, providing a high level of redundancy.
  • Pros: Very reliable, assuming only a single disk per pair fails. RAID 1 tends to provide good read performance (equal to or better than a single drive).
  • Cons: Because each drive is mirrored to another, requires 100% disk overhead to operate. Write performance can sometimes suffer due to the need to write the data to two drives, but is often still better than write performance for other RAID levels.

RAID 2: This RAID level is no longer used.

RAID 3 (Parallel transfer disks with parity)

  • Drives required (minimum): 3
  • Max capacity: (Number of disks minus 1) x capacity of each disk
  • Description: Data is broken down to the byte level and evenly striped across all of the data disks until complete. All parity information is written to a separate, dedicated disk.
  • Pros: Tolerates the loss of a single drive. Reasonable sequential write performance. Good sequential read performance.
  • Cons: Rarely used, so troubleshooting information could be sparse. Requires hardware RAID to be truly viable. RAID 3 is generally considered to be very efficient. Poor random write performance. Fair random read performance.

RAID 4 (Independent data disks with shared parity blocks)

  • Max capacity: (Number of disks minus 1) x capacity of each disk
  • Description: A file is broken down into blocks and each block is written across multiple disks, but not necessarily evenly. Like RAID 3, RAID 4 uses a separate physical disk to handle parity. Excellent choice for environments in which read rate is critical for heavy transaction volume.
  • Drives required (minimum): 3
  • Pros: Very good read rate. Tolerates the loss of a single drive.
  • Cons: Write performance is poor. Block read performance is okay.

RAID 5 (Independent access array without rotating parity)

  • Max capacity: (Number of disks - 1) x capacity of each disk
  • Description: Like RAID 4, blocks of data are written across the entire set of disks (sometimes unevenly), but in this case, the parity information is interspersed with the rest of the data.
  • Drives required (minimum): 3
  • Pros: Well supported. Tolerates the loss of a single drive.
  • Cons: Performance during a rebuild can be quite poor. Write performance is sometimes only fair due to the need to constantly update parity information.

RAID 6 (Independent Data disks with two independent distributed parity schemes)

  • Max capacity: (Number of disks - 2) x capacity of each disk
  • Description: Like RAID 4, blocks of data are written across the entire set of disks (sometimes unevenly), but in this case, the parity information is interspersed with the rest of the data.
  • Drives required (minimum): 3
  • Pros: Tolerates the loss of up to two drives. Read performance is good. Excellent for absolutely critical applications.
  • Cons: Write performance is not very good. Write performance is worse than RAID 5 due to the need to update multiple parity sets. Performance can heavily degrade during a rebuild.

No comments:

Post a Comment