Tosapol Web

tosapol@hotmail.com

[ Home ] [ Education ] [ Working ] [ Hobbies ] [ Software ] [ Hardware ] [ How-To ] [ Link ] [ mail ]

General Information about RAID

Back Ground

In 1987, professors Garth Gibson, Randy Katz, and David Patterson at the University of California outlined 6 RAID levels. These formally defined RAID levels are a family of techniques for managing a suite of disk drives so that reliability, data availability, and performance are all maximized for a wide variety of applications. Each RAID level defines how the RAID will function.

RAID Levels

Levels 0
Block Striping: RAID level 0 implements stripes data across all disks without redundancy or parity. This Level maximizes data transfer rates and is good for handling large files. Spare drives are not useful on this Level.
Levels 1
Disk Mirroring: RAID level 1 implements a mirror disk for each drive in the system. Use of a mirror drive virtually eliminates system interruption due to drive failure. Data is duplicated (in pairs) for storage on each drive. The use of dual drives more than doubles the cost of the storage unit. RAID level 1 provides fair read performance but much slower write performance since it must write to two disks.
Levels 2
Bit Interleaving: RAID level 2 interleaves bits across all drives in the array. Additional drives are inserted in the system for error correcting code or parity data. RAID 2 is typically used for large computer systems due to the robustness of its design.
Levels 3
Synchronized Spindles: RAID level 3 is similar to RAID level 2 but better suited to microcomputers. In a typical RAID level 3 system only one parity drive is used for all data drives. Interleaving of data occurs at a predefined logical size (bit, byte, block, or other unit). Each drive holds a portion of the data. The synchronized spindles yield high data transfer rates. This makes RAID level 3 ideal for high performance systems, such as image processing, where speed and accuracy are important.
Levels 4
Block Stripping with Parity Drive: RAID level 4 is RAID level 0 with a parity drive. Typically write operations are slow due to the parity read to a single drive that occurs before each write.
Levels 5
Block Striping with Distributed Parity: RAID level 5 contains both data and parity blocks on each disk, thus eliminating parity drives. Eliminating the parity drive allows the system to perform multiple read and write functions at the same time. This increases the virtual transfer rate by one-half the number of drives in the array, thus increasing performance as more drives are added.

RAID Capacity

RAID Level	Description	Theoretical Capacity	Example of Actual Capacity using seven 9 GB disks
0	Striping	# Drives x Cap Drive	63 GB
1	Mirroring	(# Drives / 2) x Cap Drive	31.5 GB
2	Hamming code parity	(# Drives -1) x Cap Drive	54 GB
3	Byte level parity	(# Drives -1) x Cap Drive	54 GB
4	Block level parity	(# Drives -1) x Cap Drive	54 GB
5	Interleave parity	(# Drives -1) x Cap Drive	54 GB

RAID Implementation

RAID can be implemented in hardware or software. Software RAID solutions use the host computer's CPU and memory to implement the RAID functions.
RAID Level 3 or 5 are usually implemented in hardware. The hardware RAID controller has a dedicated CPU to calculate parity and map the location of the files.
Internal RAID, like software RAID, is operating system dependent. It usually requires a driver to access and configure the RAID controller. On the plus side, the internal RAID controller can communicate faster than an external controller, because it incorporates the SCSI adapter function. This means that access to the data avoids one communication layer. Also, an internal RAID controller is usually less expensive than a comparably equipped external controller.
There are some major disadvantages to an internal controller.

If the controller fails, then the host computer must be shut down to repair or replace the board. If an external RAID controller fails, simply turn off the other devices on that bus.
Most internal controllers do not have expansion cards, so the size of the RAID set is limited. External controllers offer more flexibility in the number of drives it can address.
Only one host can talk to an internal controller.
Failover on an internal controller is difficult at best to configure.

RAID Features and Functions

Feature	Benefit and Advantage	Disadvantage
Hot Spare Drives	Ability to recover from single disk failure	Makes drive work but is not used
Warm Spare Drives	Ability to recover from single disk failure	Delay in recovering because the drive must be started
Global Spare Drive	Ability to recover from single disk failure on any rank	Requires multiple ranks running on the RAID controller
Failover (Active-Passive)	Avoids RAID controller as the single point of failure	Other controller sits idle until needed
Failover (Active-Active)	Avoids controller as the single point of failure	More expensive

Which is the right RAID levels?

Characteristics	RAID Level(s)
Best Cost (best utilization of disk space)	0, 3, & 5
Best Performance (Q I/O's)	0 & 5
Best Performance (data transfer)	3
Data Availability	1*, 3, & 5

*RAID Level 1 allows high availability, but requires exactly twice the needed disk space.

Courtesy of

Member of the Internet Link Exchange

Free Home Pages at GeoCities