Although there are half a dozen or so officially recognized RAID levels, and a number of combinations, there are only a few that are commonly used in small to midsized businesses (SMBs).
RAID 0 isn't really a RAID level at all because there's no redundancy. It is a shorthand way to describe striping data across several disks. It increases performance on reads and writes, but provides no additional protection against drive failure. RAID 0 is typically used to speed up the read/write performance of temporary files. With files longer than one block (approximately 4K typically), RAID 0 reads and writes from multiple disks simultaneously, speeding up access. However, because it has no redundancy, it doesn't protect the data from a disk failure.
RAID 1 refers to mirroring all writes -- writing the same information to two or more disks. The data is written twice to separate disks. RAID 1 is simple, provides excellent protection and can be restored very quickly.
The downside to RAID 1 is its cost. It is by far the most data storage-intensive method because the needed storage space is twice the size of the data being stored. So if you have 300 GB of data, you need 600 GB of disk capacity, which in turn doubles your disk costs.
RAID 3 includes parity information so data can be reconstructed in the event of a single disk failure. One disk is set aside for parity data and the data is striped across the other disks. If a disk fails, RAID 3 uses the parity data to reconstruct the contents of the failed disk.
RAID 5 is probably the most common RAID level for SMBs because it offers a good combination of protection and economy. RAID 5 stripes both the data and the parity across all the disks in the set. However, the extra overhead involved in calculating parity means it suffers a performance penalty compared to RAID 10. Not only does it take longer to write the data, but it takes longer to reconstruct the data set in the event of a disk failure. Of course, while the array is being rebuilt, it has a single point of failure in case another disk fail.
RAID 6 is RAID 5 with extra redundancy. It uses two parity disks instead of one, meaning it can recover from two failed disks. RAID 6 is becoming more common with the increased use of SATA drives, which are cheaper but less reliable than SCSI drives.
RAID 10 is an example of a nested RAID level. That is, it combines two RAID levels to gain additional benefits. RAID 10 mirrors the data across disks and then stripes the mirrored set. Striping provides higher performance and mirroring provides redundancy. RAID 10 shares the speed and simplicity of RAID 1 while providing better performance through striping. Like RAID 1, it is expensive in disk space, but many administrators are willing to pay the price.
An alternative method is RAID 01, which stripes the data across the disks and then mirrors the striped set. RAID 01 is not as robust as RAID 10 and thus isn't as popular. There are other levels. RAID 2 is never used because it has no advantages over more popular levels and RAID 4 (striping with bit rather than block parity), but it is used occasionally for applications with large sequential files.
If you need performance, RAID 10 is probably the best choice. For economy, RAID 5 gives the most redundant storage with the fewest disks. RAID 6 is useful for situations where you want the extra protection of added redundancy. In those cases, select a RAID array with more hot spares (extra disks installed and ready to go in the event of a failure).
Unless you've got a lot of disk arrays or high-performance requirements, any of the popular redundant RAID levels (i.e., anything but RAID 0) will work for SMBs.
Do you have comments on this tip? Let us know. Please let others know how useful this tip was via the rating scale below.
Do you know a helpful storage tip, timesaver or workaround? Email the editors to talk about writing for SearchSMBStorage.com.
About the author: Rick Cook specializes in writing about issues related to storage and storage management.