Related Articles

Introduction

IWill ATA100 Controller Card

IDE RAID has been around a while now, and it’s growing more and more popular among avid hardware consumers. When Iwill released the SIDE 66 card a few months ago, it proudly strutted its fast transfer speeds all over the web. But one thing stood clear –it still doesn’t match SCSI performance. But who ever said IDE and SCSI used in the same sentence would add up anyway? SCSI interface for one is much more expensive than IDE. That point alone forces many people to chose IDE instead of SCSI. Also SCSI still maintains the fastest drive performance available, making SCSI the most logical choice for high-end computers that serve an important data role.

This is where the IDE RAID comes in. It isn’t intended to replace SCSI –it’s more of an alternative. IDE RAID does have some limitations, but it also has some great benefits that rival only using one disk or even multiple software based RAID configurations.

Most of us are familiar with RAID, and the different kinds of configurations and implementations. For those who aren't too familiar with RAID, let’s do a quick rundown of RAID configs, so you will have a point of reference to return to as the review continues.

RAID overview

RAID stands for Redundant Array of Inexpensive Disks. It basically implies that two or more disks are configured in either a software based, or a hardware-based environment to serve data fault tolerance, higher data speeds, or even a combination of both.

RAID configurations can dramatically improve disk performance over the use of a single hard disk since more than one disk is doing the work. A brief definition of the different types of RAID configurations is listed below.

RAID 0 or disk striping increases the speed of the hard drive system by spreading the data writes over a disk stripe (or more than one disk). By spreading the writes over more then one drive, the data can be equally distributed using block sizes of up to 64K. The word stripe actually comes from the data writing procedure in that it cuts the data stream and only writes portions onto each disk. Since there is no parity information written along with the data, there is no fault tolerance. RAID 0 requires at least 2 disks.

RAID 1 or disk mirroring is the ability of a RAID configuration to duplicate data onto two separate drives at the same time. This comes at a slight performance cut over disk stripping since the data has to be written in two different places. At the same time, it makes the CPU work harder, but does provide fault tolerance in that if one drive fails you can quickly recover the system. Since RAID 1 uses one disk to store the duplicated drive –you will lose that area for data storage, and will be invisible to the OS. RAID 1 requires two disks.

RAID 0+1 combines the both of best worlds by stripping a mirror set. It requires 4 –but this will provide you with fault tolerance plus performance. This is one of the most costly RAID configurations because there are two drives that have to be replicated, the other two are needed for data redundancy –making drive real-estate increase immensely!

RAID 5 is primarily used in higher end SCSI configurations. It provides speed and fault tolerance, (only in a single drive failure). In a RAID 5, the data is stripped across all the disks in the configuration along with parity information. Since the equivalent in space to one disk is used to store the parity information (to recover from a drive loss) you will lose that area for data storage. RAID 5 can also use many more disks than other RAID configs, making this the clear choice when mass storage is needed.

send article   hardware newsletter   article comments (12)

Submit our article to: Digg this! del.icio.us