Promise FastTrack/FastSwap IDE RAID Controllerby Anand Lal Shimpi on March 16, 1999 11:55 AM EST
- Posted in
What is RAID?
The first thing that will pop into most minds after hearing the name "RAID" is the common household bug spray. And if you're a member of Promise Technology's marketing team, the last thing you want is a product that has the tone of a household bug spray ringing in the minds of those that see it. At the same time, there is a definite lack of education on the topic of the other meaning for RAID both online and off. If you were to make two separate posts on any general computer hardware news site online today, one concerning the latest in SLI technology for gamers and the other dealing with a new RAID controller card, chances are that the number of clicks on the SLI news blurb would be almost triple that of the RAID blurb. Why? Well, why don't you ask yourself, how many users out there do you really think know about the true functionality and benefits of RAID? Then take that number and compare it to the number of users out there have any idea about the benefits of a SLI video solution after 3Dfx's eruption with the whole Voodoo2 SLI extravaganza in 1998? What you'll probably come up with is a definitely uneven ratio. So what's the first step in understanding Promise's IDE RAID Technology? Explaining the acronym RAID of course.
RAID, generally speaking, is a technology that allows two independent disks, or two separate hard drives to function as a single drive equal in size to the capacity of the smallest drive in the collection of drives (array) multiplied by the number of drives. What this means is that if you were to have a RAID setup of 4 drives, 3 of which happened to be 10GB drives and one which was a 1GB drive, the total capacity of the RAID system would be 4GB, or the size of the smallest drive in the array multiplied by the number of drives in the array (1GB x 4 drives = 4GB). The purpose of RAID can be either to improve performance, increase reliability, or both.
RAID 0 - Striping
Imagine yourself sitting in front of two bowls of peanuts at a party. Now you happen to be very hungry, and want to grab as many peanuts before the hungry guest no one invited makes his way over to the bowls and begins to hog all of the snacks. You could use a single hand to grab a handful of the peanuts from one of the bowls and move on, or you could take another approach. Couldn't you, theoretically, use both hands, reaching in to both identical bowls of peanuts, and grab twice as many in the same amount of peanuts? Absolutely. This is the same principle that is behind one function of RAID, known as striping.
In order to improve performance, striping is used. Striping is a function of RAID, known as RAID 0, that allows for data to be read off of different drives in the array at the same time instead of going to a single drive for all of the data. For example, if you were reading 2 individual sectors of a file that was to be loaded, the standard way of proceeding with such a process would be that the two sectors would be read in sequence from your single hard drive. However, in a RAID 0 capable system with a disk array consisting of two drives, the first sector would be read off of the first drive while the second sector would be read off of the second drive, at the same time, therefore doubling the theoretical performance of the data transfer. Suppose your disk array consisted of four drives, and 4 sectors were to be read, the same would apply, where the first drive would read the first sector, the second would read the second sector, the third would read the third sector, and finally the fourth drive would read the fourth sector. Make sense? That basic technology, known as RAID 0, or striping, is how a RAID system can improve performance. How about reliability?