IN starting this article; I’m going to assume that you know what RAID is: If not then a click on the link on the word RAID in the previous sentence will take you to Wikipedia where you can ‘read about RAID’.

Using the RAID 0 (RAID zero) configuration requires the implementation 2 or more drives, on which data is shared across these drives in the array. This means that each drive gets roughly the same amount of data written to it, and therefore the drive-access speed is substantially increased. – So far it sounds like an excellent idea!

…But there is just one drawback; one huge drawback: If one of the two or more drives in the RAID 0 array fails, then all data on all drives is lost.

“Right… That happens with a single drive too:” You interject. “If it fails then all data is lost. – What’s the difference?”

The difference is; dear reader, that each drive that you add to the RAID 0 array increases the likelihood of one of the drives failing: -

You start off with a single drive. This drive has an approximate expected lifespan of 3 years (No leap-years.), or 1095 days. – Therefore the projected likelihood of the drive failing on any given day, ignoring the exponential increased possibility of failure with use and age, is 1 in 1095.

You then decide to utilise RAID 0; so you add a second, identical drive, and create an array: You now have, in effect, a hard-drive with double the access-speed. – You have just increased the chances of disc-failure causing total data loss to 2 in 1095: Just under 1 in 550.

Adding a third, identical drive, will increase the disc-access speed still further; yet it will also increase the chances of failure and total data loss to 3 in 1095, or 1 in 365: From that number you can assume that the 3-drive-array will fail every year, and unless you take regular backups you get screwed once a year. (A bit like my sex-life; but that’s enough exaggeration for now.)( – I’m a geekette: I’d rather get my kicks from technology. – At the end of the day it’s still a lot less complicated, and leaves no physical and emotional scars as such when it goes wrong.( Buzz, buzz…phut!.. !@!*!))

- I can read your mind!

“But it won’t fail once a year in reality, will it! – Come on Sharron; you’re intelligent enough to be able to work out that it’s extremely unlikely that there will be any failures in the first year, and probably none in the second year either.”

I did state that I was ignoring the exponential, in other words speaking and thinking linearly, for the sake of convenience. – OK smartass, let’s complicate it: Cast your mind back to September 2009, and the article I wrote entitled “Hard-Drives Are Unreliable – Fact ". One disc in 14 in the Google analysis failed after 2 years; which gives a 1 in 6.33 chance of any individual disc in a 3-disc-array failing after 2 years: In other words there’s a 49/51 chance that one disc will fail at two years old.

Of course that study was done on top-quality industrial hard-drives that were built to last; and in said article we worked out that the average life of a disc used in the study was 6 years. In this article I’m being a bit more realistic by using half that figure as the average life of one of the discs in the array; which is nearer the expected lifetime of a cheaper commercial-sector disc-drive, such as a Maxtor, for instance. On the basis of that correction we can say that there’s a 49/51 chance of one of the drives failing after the first year by logical reasoning, or a 75% chance of a failure within the first 2 years: In other words, although there’s nothing definite there; it’s a lot more likely than not that something pops off in the first 2 years. – In summary; you should expect a drive in a 3-drive RAID 0 array to die after 2 years of operation, if it hasn’t done so already.

‘Happy now?

- So RAID 0 is great for vastly improving disc-access times; but don’t expect it to last long.

RAID 5, on the other hand, is a much better bet, will give a similar performance-boost to RAID 0, and won’t trash all your data if a drive fails.

RAID 5 requires 3 or more drives, and retains parity information, so that 99 times out of 100 the lost data can be recalculated if a single drive fails… Is that another article’s worth? – Yes; that’s for another article.

Use RAID 0 at your own peril; and if you do use it regardless, then I suggest that you replace all of the drives in the array if one fails, and that you backup your data at least once a day. The question you have to ask yourself is this: Is three or more times the expense worth it for the extra performance-speed? Time is money. – Either extra expense shaving seconds off it, or costing you in lost performance if you don’t.

It’s a bit “heads time wins; tails you lose” if you ask me.

How do you perceive it?