RAID Levels

Redundant array of inexpensive disks (RAID) is used to configure a disk subsystem to provide better performance and fault tolerance for an application. The basic idea behind using RAID is that you spread data across multiple disk drives so that I/Os are spread across multiple drives. RAID has special significance for database-related applications, where you want to spread random I/Os (data changes) and sequential I/Os (for the transaction log) across different disk subsystems to minimize disk head movement and maximize I/O performance.
Since multiple disks increases the mean time between failures (MTBF), storing data redundantly also increases fault tolerance. A RAID appears to the operating system to be a single logical hard disk. RAID employs the technique of disk striping, which involves partitioning each drive's storage space into units ranging from a sector (512 bytes) up to several megabytes.
RAID levels

RAID-0
RAID-0 uses disk striping; that is, it writes data across multiple hard disk partitions in what is called A stripe set. This can greatly improve speed because multiple hard disks are working at the same time. You can implement RAID-0 through the use of Windows Server software or third-party hardware. Although RAID-0 gives you the best speed, it does not provide any fault-tolerance. If one of the hard disks in the stripe set is damaged, you lose all of your data. Because of the lack of faulttolerance, Microsoft doesn’t recommend storing any of your SQL Server data on RAID-0 volumes.

RAID-1
RAID-1 uses disk mirroring. Disk mirroring actually writes your information to disk twice — once to the primary file and once to the mirror. This gives you excellent fault-tolerance, but it is fairly slow, because you must write to disk twice. Windows Server allows you to implement RAID-1 with a single controller, or you can use a controller for each drive in the mirror, commonly referred to as disk duplexing. This is the recommended place for storing your transaction logs because RAID-1 gives fast sequential write speed (writing data in sequence on the disk rather than jumping from one empty spot to the next), a requirement for transaction logs.

RAID-5
RAID-5—striping with parity — writes data to the hard disk in stripe sets. Parity checksums will be written across all disks in the stripe set. This gives you excellent fault-tolerance as well as excellent speed with a reasonable amount of overhead. You can use the parity checksums to re-create information lost if a single disk in the stripe set fails. If more than one disk in the stripe set fails, however, you will lose all your data. Although Windows Server supports RAID-5 in a software implementation, a hardware implementation is faster and more reliable, and we suggest you use it if you can afford it. Microsoft recommends storing your data files on this type of RAID because data files require fast read speed as opposed to transaction logs, which need fast write speed.

RAID-10
You should use RAID-10 (sometimes referred to as RAID 0+1) in mission-critical systems that require 24/7 uptime and the fastest possible access. RAID-10 implements striping with parity as in RAID-5 and then mirrors the stripe sets. So, you get the incredible speed and faulttolerance, but RAID-10 has a drawback. With this type of RAID you get the added expense of using more than twice the disk space of RAID-1. Then again, we are talking about a situation that can afford no SQL Server downtime
Unless you can afford a RAID-10 array, Microsoft suggests a combination of RAID-5 and RAID-1. In this scenario, you place your data files on the RAID-5 array for speed and redundancy. You place your transaction log files on the RAID-1 drives so they can be mirrored.