RAID Basics

RAID stands for Redundant Array of Independent Disks. There is two main implementations of RAID; software and hardware RAID. In my experience I have found that most Information Technology professionals steer well away from software RAID, in fact that hate it with a passion. This is mainly because of the CPU and other system overheads and the stability of software. The fact that hardware RAID controllers and onboard RAID controllers are becoming very inexpensive there is decreasing reasons why anyone would use software RAID, like that implemented in Microsoft Windows 2000pro/XPpro. This said there is also many advantages to software RAID, but in my, and others opinions these are not worth the losses.

RAID attempts to address many issues within the computer system storage sector. The first of these is mass storage. Raid can daisy chain many disks together and appear to the user and O/S (operating system) as a single disk and single or multiple partitions. One off the major bottlenecks in computer systems today is fast storage devices. Hard disk drives are mechanical devices, unlike storage such as RAM

The first type of RAID is RAID level 0. People refer to a RAID0 implementation as striping. A simple example of how this works is as follows: Imagine you are filling a bath tub with water to a certain level from a small tap that’s on as fast as it can go. The bath tub will be full to that level in x amount of time. If we want to fill the tub faster we can add a second tap and it will theoretically fill twice as fast. Now assume you have a 500MB file (water) you have to write to your 40GB hard disk drive. The file will take x time to write to disk because the disk can go so fast but not any faster (tap size). Now instead of writing the 500MB file to a single disk imagine if we wrote half of the file (250MB) to two (add a second tap) 20BG Hard disks. Theoretically it should also be tow times as quick! This is how RAID increases performance. Chunks (blocks) of KB are written to each disk. So now the minimum time taken to write 500MB to one disk is halved because only 250MB has to be written to one disk. If we add yet another disk (another tap) only ~166MB needs to be written to each disk and so we add more disks. It is ideal to use the same type and size of hard disk when striping. If different sizes are used the total RAID0 capacity is no of drives in array times smallest disk capacity. We should avoid wasted capacity. RAID0 adds no redundancy. RAID0 is mainly used is performance critical applications where redundancy is not required or is implemented via another means. It is becoming popular is small office and home PC’s.

RAID0/1 or RAID10 uses both the above implementations at once. Suppose we use 2 disks for striping (RAID0) and we take a ‘mirror’ of this using 2 more disks! This is classed as RAID10 or 0/1 and takes a minimum of four hard disks to operate. T

Some topics in this essay:
PC’s RAID, RAID RAID, RAID4 RAID, Flash ROM, Microsoft Windows, RAID0/1 RAID10, Information Technology, raid level, Independent Disks, , hard disk, 2 disks, single disk, written disk, disk add, software raid, hard disks, disk 2, 2 disks striping, 2 parity, raid level 4, mass storage raid, software hardware raid, written disk add,