RAID (Redundant Array of Independent Disk)

RAID (Redundant Array of Independent Dsik)

Raid Card Controler (Add-in)

In computers, RAID refers to data storage scheme using multiple hard drives to share or replicate data between drives. Depending on the configuration of the RAID (typically referred to as RAID levels), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In the original implementation, the main advantage is the ability to combine several low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than is available in one affordable device using the latest technology.
In other words RAID function is to merge several hard drives into one logical unit or a single volume.

History

In 1978, Norman Ken Ouchi from International Business Machines (IBM) was awarded a U.S. patent, number 4,092,732, titled "System for recovering data stored in failed memory unit." This patent claims to explain about what became known as RAID 5 with the writing of the full stripe. Patent in 1978 that also mentions that disk mirroring or duplexing (now known as RAID 1) and also special protection with dedicated parity (which is now known as RAID 4) can be used, although at that time there has been no implementation.
The term "RAID" was first defined by David A. Patterson, Garth A. Gibson and Randy Katz at the University of California, Berkeley, USA in 1987, 9 years ago after a patent owned by Norman Ken Ouchi. The three of them learn about the possible use of two or more hard disks to be seen as a single perangat by systems that use them, and then they publish it in the form of a paper entitled "A Case for Redundant Arrays of Inexpensive Disks (RAID)" in June 1988 on SIGMOD conference time. The specification offers some prototypes RAID level, or a combination of these drives. Each RAID level is theoretically has advantages and disadvantages of each. One year later, RAID implementations began to surface a lot. Most of these implementations are substantially different from the original RAID levels made by Patterson and his colleagues, but the implementation uses the same number with what is written by Patterson. This can be confusing, as an example of one implementation of RAID 5 to RAID 5 implementations differ from each other. RAID 3 and RAID 4 also can be confusing and often interchangeable, although basically these two different types of RAID.



Raid Levels
1. RAID 0
Also known as stripping mode. Requires at least 2 hard drives. The system is to combine the capacity from multiple disk. Thus, it is logical only "seen" a hard drive with large capacity (number of total disk capacity).


At first, RAID 0, used to form a partition that is larger than some hard drive in a cost-efficient.


For example:
We need a partition of 500GB. The price of a 100GB hard drive size is Rp.500.000, - while the price of a 500GB hard drive size is 5.000.000, -. Well, we can set up a 500GB one partition size of 5 units sized 100GB hard drives using RAID 0. Of course, this scenario is less expensive because of the cost is cheaper: 5 x Rp.500.000, - = Rp.2.500.000, -. It's cheaper than having to buy a 500GB hard drive size. That is why at first called the Redundant array of Inexpensive disks.


Another example:
At the current size of the largest hard drive available on the market is 500GB, while we need a partition of 2TB. Well, we can buy 4 units of 500GB hard drive and configured with RAID 0, so that we can have a partition without having to wait berkururan 2TB hard drive with that much capacity available in the market.


Data is written on the disk-drive is divided into fragments. Where fragments are scattered throughout the hard disk. Thus, if one hard drive suffered physical damage, then the data can not be read at all.
But there are advantages in the presence of these fragments: speed. Data can be accessed much faster with RAID 0, because when the computer reads a fragment on one disk, the computer can also read another fragment in the other hard drive.

2. RAID 1

     Usually called mirroring mode. Requires at least 2 hard drives. The system is copying the contents of a hard drive to another hard drive with a purpose: if one hard drive is physically damaged, then the data remains accessible from other hard drive.


Example:
      A server has 2 hard drive unit with a capacity of 80GB each and configured RAID 1. After several years, one of the hard drive suffered physical damage. However, other data on the hard drive can still be read, so that data can still be saved as long as not all physically damaged hard drive simultaneously.

3. RAID 2

      RAID 2, also using a stripping system. However, added three more to pariti Hamming hard drive, so the data becomes more reliable. Therefore, the number of hard drives required is less than 5 (n +3, n> 1). Third last hard drive used to store the Hamming code in the calculation of each of bits that exist in other hard disk.


Example:
     We have 5 hard drive (say drive A, B, C, D, and E) with the same size, each 40GB. If we set the four hard drives with RAID 2, then the capacity obtained is: 2 x 40GB = 80GB (of hard drive A and B). While hard drives C, D, and E are not used for data storage, but only to store information pariti Hamming from two other hard drives: A and B. When there is physical damage to one of the main hard drive (A or B), then the data can still be read with consideration pariti Hamming code on the hard drive C, D, and E.

4. RAID 3

     RAID 3, also uses the stripping system. Also use an additional hard drive for reliability, but only added a hard drive for parity again .. Therefore, the number of hard drives required is a minimum of 3 (n +1, n> 1). Last hard drive used to store the parity of the calculation of each of bits that is on another hard drive.


Case in point:
     We have 4 hard drive (say drive A, B, C, and D) with the same size, each 40GB. If we set the four hard drives with RAID 3, the capacity obtained is: 3 x 40GB = 120GB. While the disk D is not used for data storage, but only to store parity information from the three other hard drives: A, B, and C. When there is physical damage to one of the main hard drive (A, B, or C), then the data can still be read by taking into account that there is parity in the disk D. However, if the disk D that were damaged, then the data can still be read from the three other hard disk.

5. Raid 4

   Similar to RAID 3 systems, but using a parity block from each disk, not a bit. Minimum hard disk requirement is the same, 3 (n +1, n> 1).

6. RAID 5

     RAID 5 is basically the same as RAID 4, but with a distributed pariti. Namely, do not use a special hard drive for storing paritinya, but paritinya spread to the entire hard drive. Minimum hard disk requirement is the same, 3 (n +1, n> 1).


     This is done to speed access and avoid the bottleneck that occurs because disk access is not focused on the collection of hard drives that contain data only.

7. RAID 6

    In general is an improvement over RAID 5, namely with the addition of parity into 2 (p + q). So the minimum number of hard drives is 4 (n +2, n> 1). With the addition of this secondary pariti, then the damage to two hard drives at the same time can still be tolerated. For example, if a damaged hard drive, hard drive during the exchange process occurs more damage at one another hard drive, then it can still be tolerated and not cause data loss in disk-system RAID 6.

Anda baru saja membaca artikel yang berkategori Kuliah / Pelajaran dengan judul RAID (Redundant Array of Independent Disk). Anda bisa bookmark halaman ini dengan URL http://link-soft.blogspot.com/2011/05/raid-redundant-array-of-independent.html. Terima kasih!

Ditulis oleh: Unknown - Tuesday, May 3, 2011