When we hear the word RAID we may cringe at the sound of it. To be honest it just sounds mean. But it really is nothing to be afraid of. In fact, putting hard drives into arrays in order to gain more performance, more storage space, data protection, or all of the above can not only speed up your workflow, it can one day save you in the event you have a hard drive failure.
So what is this RAID you speak of? I thought I used that to kill ants!
RAID stands for “Redundant Array of Inexpensive Disks”. (I didn’t name it so no e-mails about how it’s not necessarily that inexpensive please.) What it does, is strap multiple hard drives together and presents it to your computer as one large volume. In a RAID, data is distributed between several physical disks, either called member disks or raid slices, to provide more reliability and/or speed increase. There are multiple RAID “levels” which differ in the way that they read, write and store data. The three most popular of which will be explored more below.
RAID can be utilized either by a driver in the operating system (Software RAID) or by a special hardware controller (Hardware RAID). Both OSX and Windows Vista provide their own software RAID implementations. As robust as this may be, be aware that a RAID is not a substitute for a proper and regular backup and archiving plan. Multiple drive failures can and do happen, also if the controller fails the entire array can be lost. Nevertheless, a RAID solution might just be what the doctor ordered if you need additional faster storage.
By far, the most popular RAID level is RAID 0. This is a “non-redundant” RAID. This means that if any one drive that belongs to the array fails, the whole array is lost. The most typical use of this level of RAID is to gain performance. The way this is accomplished is by scattering the read and write requests evenly across member disks. A minimum of two disks is required for this configuration, and there is no disk space overhead required.
RAID 1 or “Mirrored” RAID as it is sometimes called is just that. Two exact copies of data are written on two member disks. The second disk, or “mirrored” is an exact copy of its primary disk. This configuration can withstand a loss of any single disk. Speed with this level of RAID is not improved because all copies of the mirror need to be updated. A minimum of two disks are needed for RAID 1 and there is a disk space overhead of 100%.
RAID 5 provides redundancy and data reconstruction. What this means is that in case of a single drive failure the drive can be replaced and rebuilt into the array without loss of data. During normal operation read speed is very close to RAID 0, because the data can be read evenly across the disks. Write speed however is somewhat diminished as the data must not only be written to the array but also to a parity disk. (thus the redundancy). This can be mitigated however by raising the number of disks in the array. A minimum of three disks id required for RAID 5 and the disk space overhead equals the capacity of one single member disk.
So now that you have a little more understanding of RAID and how it can help you, hopefully you can move past the voodoo involved with adding storage and deciding which will be best for you.