|
|
|
|
 |
EIDE RAID Iain Laskey explains EIDE RAID and looks at what it can do for you. RAID stands for Redundant Array of Independent Drives. The acronym originally stood for Inexpensive Drives but given that the technology was originally extremely expensive, this seemed a strange way to describe it. Until recently all RAID systems were built using SCSI drives and specialised drive controllers. Because of their high prices they tended to be reserved for high performance systems such as large network servers or database servers where performance was far more important than price. RAID was firmly in the SCSI domain purely because EIDE (commonly referred to as just IDE) drives did not have the required performance. Now that we have ATA-66 and more recently ATA-100 drives, RAID using EIDE is now possible. There are two options to get you RAIDed up. Firstly you can buy motherboards with built in EIDE RAID controllers. These can be either the only EIDE drive controllers or in addition to the usual four EIDE drives you can connect. Another option is to buy an EIDE RAID controller for plugging in to a PCI slot. These can be as cheap as £40 or so but do almost as much as their SCSI counterparts that often cost well over £1000. To confuse things, there are several flavours of RAID depending on what you wish to achieve but essentially they all involve two or more hard drives being used together to achieve faster throughput or data security. EIDE RAID has become even more compelling of late due to the massive reduction in hard disk prices with 40Gb drives easily picked up for £75. Let’s look at the more common types of RAID to see what they can do for you. RAID 0 RAID 0 is called striping and treats two or more drives and one large drive. For instance, you could install two 40Gb drives configures as RAID 0. Your PC would actually see this as one 80Gb drive going almost twice as fast as a single drive because data is read from both physical drives simultaneously. If you installed four 40Gb drives you would get a 160Gb drive going four times as fast as a single drive. Strictly speaking, this is more of an AID system as there is no redundancy here. We’ll see an example of redundancy in action under RAID 1. The drives under RAID 0 should ideally be the same size but you can mix different sizes. The downside is that the system will treat all drives as if they were the size of the smallest drive. As an example, a 40Gb, 20Gb and 15Gb combination would be seen as a 45Gb (3 X 15Gb) system but still getting the 3 X speed improvement. RAID 1 RAID 1 is known as mirroring. Here the RAID controller keeps identical copies of data on each drive. This is where the redundancy comes in. A RAID 1 system could use 2 x 40Gb drives to give one 40Gb drive with redundancy. If a drive fails, you won’t lose any data as long as you replace it before the second one fails. When the new drive is installed, the controller copies all the data from the existing drive to ensure continued redundancy. As a user, you need to do nothing. You won’t get a speed boost via RAID 1 though, it’s all about redundancy. RAID 0+1 You can combine RAID 0 and 1 to get the best of both worlds. Here you could use 4 X 40Gb drives to get a single 80Gb drive with redundancy. If any drive fails, simply replace it and continue working. You’ll have no loss of data or need to restore as two of the drives will be holding a mirror image of the other two. Additionally, you will get double the throughput because each pair of drives will be working in RAID 0 mode. RAID 5 RAID 5 is the deluxe option and not yet very common with EIDE RAID systems which generally only support 0, 1 and 0+1. RAID 5 needs a minimum of three drives but can have as many as eight or more on bigger systems. RAID 5 looks at the data on all the drives bar one and uses that to calculate a checksum which is stored on the final drive. If any single drive fails (including the checksum drive) it can be rebuilt by recalculating the data based on the checksums. RAID 5 maximises the value of the drives by giving the speed of many drives reading and writing simultaneously combined with the redundancy of RAID 1. Setting It Up Now we know what RAID is, how is it set up? RAID controllers are normally accessed via a configuration screen not unlike your motherboard’s BIOS screen. From here you can set the type of RAID and how the drives are to be used. Once this is done, the PC will no longer see several different drives but just see the one drive you have created. This means it is transparent to Windows, no special drivers or setting up is needed. Why? RAID isn’t for everyone. With modern ATA-100 drives capable of running at 20Gb/second or more, you’d think that would be enough for everyone. However, applications that throw vast amount of data around such as photo editing, or audio processing can benefit. Also, the redundancy aspect can be useful if you wish to protect important data at all times without worrying about doing backups. RAID can also be effective for video editing systems although recently some people have found that capturing video can result in lost frames as the PCI bus gets saturated feeding the RAID system with data whilst giving the video capture card the bandwidth it needs to pull real-time video in to the system. That said, I know of several people using RAID 1 and RAID 0+1 systems to edit video on with no problems whatsoever.
|
