RAID

Redundant Array of Inexpensive Disks

The Basics

A simple definition of a RAID system is an array of disks functioning as a single unit.

The objective of RAID systems is to provide and large amount of disk storage for a low cost while allowing for a set number of disk hardware failures.

For the most part, RAID falls into two implementations, hardware and software.

The models used for writing the data to the disk and the number of hardware failures sustainable without data loss are defined as levels.  The original definitions for the RAID levels came from a Berkley paper title "A Case for Redundant Arrays of Inexpensive Disks".  Six levels were defined by the writers, but more have been suggest by others. 

All of this is information that every sales person spouts constantly.  When you ask most people, they will tell you that it requires one extra disk, that is spread around the data so that you can lose a disk, or my favorite it uses a parity disk to save the information of the other disks.  All of these fall short of telling how one disk can contain enough information to rebuild all of the others and what parity actually is.

Other Points

Levels

Implementation

Parity

Q&A

Other Sites

http://www.oreilly.com/reference/dictionary/ter.../Redundant_Array_of_Inexpensive_Disks.htm

http://www.busn.ucok.edu/tips/info_hrd/raid.htm

http://www.eurologic.com/tn/tnwp3.htm

http://www.raid-storage.com/raid.html

This page brought to you by James Crifasi for CMSC 411

mailto:jcrifasi@erols.com