• Choose a category
• All
• Classic-Fiction

The one and only advantage of disk striping is speed—it is a fast way to read and write to hard drives. But if either drive fails, all data is lost. You should not do disk striping—unless you’re willing to increase the risk of losing data to increase the speed at which your hard drives save and restore data.

Figure 11-35 Disk striping

Disk striping with parity, in contrast, protects data by adding extra information, called parity data, that can be used to rebuild data if one of the drives fails. Disk striping with parity requires at least three drives, but it is common to use more than three. Disk striping with parity combines the best of disk mirroring and plain disk striping. It protects data and is quite fast. The majority of network servers use a type of disk striping with parity.

RAID

A couple of sharp guys in Berkeley back in the 1980s organized the many techniques for using multiple drives for data protection and increasing speeds as the redundant array of independent (or inexpensive) disks (RAID). They outlined seven levels of RAID, numbered 0 through 6.

* * *

NOTE An array in the context of RAID refers to a collection of two or more hard drives.

RAID 0—Disk Striping Disk striping requires at least two drives. It does not provide redundancy to data. If any one drive fails, all data is lost.

RAID 1—Disk Mirroring/Duplexing RAID 1 arrays require at least two hard drives, although they also work with any even number of drives. RAID 1 is the ultimate in safety, but you lose storage space because the data is duplicated; you need two 100-GB drives to store 100 GB of data.

RAID 2—Disk Striping with Multiple Parity Drives RAID 2 was a weird RAID idea that never saw practical use. Unused, ignore it.

RAID 3 and 4—Disk Striping with Dedicated Parity RAID 3 and 4 combined dedicated data drives with dedicated parity drives. The differences between the two are trivial. Unlike RAID 2, these versions did see some use in the real world but were quickly replaced by RAID 5.

RAID 5—Disk Striping with Distributed Parity Instead of dedicated data and parity drives, RAID 5 distributes data and parity information evenly across all drives. This is the fastest way to provide data redundancy. RAID 5 is by far the most common RAID implementation and requires at least three drives. RAID 5 arrays effectively use one drive’s worth of space for parity. If, for example, you have three 200-GB drives, your total storage capacity is 400 GB. If you have four 200-GB drives, your total capacity is 600 GB.

RAID 6—Disk Striping with Extra Parity If you lose a hard drive in a RAID 5 array, your data is at great risk until you replace the bad hard drive and rebuild the array. RAID 6 is RAID 5 with extra parity information. RAID 6 needs at least five drives, but in exchange you can lose up to two drives at the same time. RAID 6 is gaining in popularity for those willing to use larger arrays.

* * *

NOTE No tech worth her salt says such things as “We’re implementing disk striping with parity.” Use the RAID level. Say, “We’re implementing RAID 5.” It’s more accurate and very impressive to the folks in the accounting department!

After these first RAID levels were defined, some manufacturers came up with ways to combine different RAIDs. For example, what if you took two pairs of striped drives and mirrored the pairs? You would get what is called RAID 0+1. Or what if (read this carefully now) you took two pairs of mirrored drives and striped the pairs? You then get what we call RAID 1+0 or what is often called RAID 10. Combinations of different types of single RAID are called multiple RAID or nested RAID solutions.

* * *

NOTE There is actually a term for a storage system composed of multiple independent disks rather than disks organized by using RAID: JBOD, which stands for just a bunch of disks (or drives).

Implementing RAID

RAID levels describe different methods of providing data redundancy or enhancing the speed