Data Availability: RAID

Single RAID Levels

For each of the RAID levels discussed on this page, a demonstration of the RAID level's method of data distribution is included. Click on the Play button to view the demonstration; click on Reset to reset the animation.

You will need Macromedia's Flash Player^TM to view the animations. Netscape users should click here to download a Netscape-specific version of the Flash Player.

RAID 0 - Disk Striping   RAID level 0 is also known as disk striping. Data is broken down into segments. Each segment is written sequentially, one segment at a time: first to one disk, then to the next disk, and so on until all of the data has been distributed among all hard disks. Disk I/O is greatly improved, because the data is stored across multiple disks and the data segments can be read almost simultaneously. RAID 0 improves disk performance considerably.   However, there is no data redundancy with RAID 0. If one drive fails, all the data is lost. A RAID 0 implementation requires at least two hard disks.   Advantages Limitations   Simple to implement Better read and write performance than a single disk Minimum of two hard disks are needed   No data redundancy All data lost if one drive fails Not recommended for mission critical environments

RAID 1 - Mirroring   RAID level 1 is also known as disk mirroring. Data is written to two disks simultaneously, providing complete redundancy. Read performance is not affected, but write performance suffers a bit because data must be written twice (once to each disk). The mirror's capacity is essentially the capacity of one disk. If you have two 18GB hard disks configured in a RAID 1 mirror, your usable capacity is 18GB. The data contained on one of the hard disks is duplicated to the other.   A minimum of two hard disks are required to implement disk mirroring.   Advantages Limitations   Simple to implement Better read performance than a single disk Good rebuild performance One disk can fail without data loss   Slightly slower write performance than a single disk More "expensive" since twice the number of disks are used than are required to store the data without redundancy

For the RAID 3 and RAID 5 animations below, clicking the Parity button will demonstrate how the RAID level distributes its data and generates its parity information. The Rebuild button will show the process by which the RAID level reconstructs data lost due to a failure.

RAID 3 - Disk Striping with Dedicated Parity Disk   RAID level 3 utilizes disk striping with a dedicated parity disk. RAID 3 provides data redundancy by using data striping in combination with parity information. Data is striped across the array disks, with one disk dedicated to parity information. If a drive fails (as in Step 4 of the walkthrough), the data can be reconstructed from the parity information stored on the parity disk (Step 5). A RAID 3 implementation requires at least 3 hard disks.     Advantages Limitations   Simple to implement Better read performance than a single disk One hard disk can fail without data loss   Slower write performance than a single disk RAID array operates in a degraded state upon drive failure Fair rebuild performance

RAID 5 - Disk Striping with Distributed Parity   RAID level 5 is known as disk striping with distributed parity. Data is striped in the same manner as RAID 3; however, RAID 5 distributes parity information across all the disks in the RAID array. If a drive fails, the data can be reconstructed from the parity information stored across all the disks. A RAID 5 implementation requires at least three hard disks.     Advantages Limitations   Simple to implement Better read performance than a single disk One hard disk can fail without data loss   Slower write performance than a single disk RAID array operates in a degraded state upon drive failure Poor rebuild performance