RAID

More about RAID

RAID technology has been a major boon to server administrators everywhere. It has allowed for seamless data redundancy and greater performance by using existing hardware technology. But what is RAID? How does it work? Why is it used so extensively?

RAID stands for 'Redundant Array of Independent Disks.' Some groups use the term 'Inexpensive' instead of 'Independent.' Primarily, it is used to spread data across multiple hard disks to provide data redundancy (in case one of the drives fails) and to spread data throughout several disks (to allow for the data to be retrieved at higher speeds). It can also be used to create one large virtual hard drive from several smaller drives.

One of the main developments which has boosted interest in RAID is the introduction of low cost RAID controllers. Originally, RAID was only usable with SCSI drives. While SCSI drives are definitely well suited to this sort of application, they are a great deal more expensive than IDE drives and the cost of multiple SCSI drives left RAID outside of the price range for many small companies and independent server administrators. However, with low cost controllers available which allowed the use of IDE drives, the cost of setting up and maintaining a RAID array dropped dramatically.

There are numerous 'flavors' of RAID called 'levels.' Each RAID level provides a different service for the machine that it is installed on. One common misconception is that since they are called 'levels,' RAID level N isn't as good as RAID level N+1. This is not the case at all. It is not a hierarchical rating, merely a title.

RAID was originally introduced in 1988. When it was released, only levels 1-5 were defined. Levels 0 and 6 were introduced afterwards, and other proprietary levels, such as RAID level 7, were developed independently by a variety of companies.

Some forms of RAID are the sole property of the companies that developed them. A majority, however, are usable by everyone. Three of the most popular implementations of RAID are level 0, level 1, and level 4.

Level 0 RAID is also known as 'Striped Disk Array without Fault Tolerance.' It is used exclusively to spread data between two or more hard drives. This provides a higher level of performance, since data can be pulled from multiple disks simultaneously. The normal bottleneck with retrieving data from a hard disk involves the speed of the hard disk. If you had a level 0 RAID array set up with 4 disks, data could be pulled from the single virtual disk it creates at up to 4 times the speed of a single disk of the same size. The main disadvantage of a level 0 RAID array is that there is no fault tolerance. If one of the disks in the array fails or is corrupted, all of the data will be lost.

The data is saved to the multiple drives through a method called data striping, where blocks of data are saved sequentially to the drives. For example, in a 3 drive level 0 RAID array, block A would be saved to drive A, block B to drive B, block C to drive C, block D to drive A, and so on until all of the data was written to a drive. The data is written and read from the drives simultaneously, meaning blocks A, B, and C would all be written/read at the same time before moving on to blocks D, E, and F. Three times the data would be written or read in the same time a normal drive would take.

Level 1 RAID arrays are also known as 'Mirroring and Duplexing.' Though they are both forms of the same technology, level 1 RAID arrays provide none of the same benefits of level 0 arrays. While level 0 arrays are a way to increase disk performance, level 1 arrays give servers redundancy. Normally, data saved to a hard drive is written to one disk and that's the end of the story. With a level 1 RAID array, the data is written to every drive in the array simultaneously.

While level 1 RAID does not provide any additional speed or performance, if there is a disk failure, the exact same data is available on another disk, which means that there is full redundancy. There are many fewer concerns with data recovery using a level 1 RAID array, which means that is adds an additional layer of protection for sensitive data.

Level 4 RAID is also known as ' Dedicated Parity Drive .' It has many of the advantages of both level 0 and level 1 RAID by using data striping to share information between several drives and also making a full copy of the data onto a backup or 'parity' drive. The data is therefore safe from disk failure and can also be read quickly. The only problem that level 4 RAID arrays have is that there is a bottleneck when writing data since the parity drive does not benefit from the performance enhancement of data striping.

Each type of RAID has its own inherent advantages and disadvantages. As a means of overcoming these limitations, level 0+1 RAID was created. Level 0+1 RAID is also known as a 'Mirror of Stripes.' With this version of RAID, two data stripes are created (much like in level 0 RAID) and a mirror is created over them (like in level 1 RAID). This means that the performance increase and the data redundancy are both preserved.

RAID 0+1's sole disadvantage is that only two drives can be used in the array. This seriously limits the performance and size benefits that level 0 RAID provides. However, the mirroring no longer causes a bottleneck, like it does in level 4 RAID arrays. By placing a limit on how beneficial the array is, level 0+1 RAID allows for all of the benefits to be used simultaneously.

Though the variations listed above are the most popular, they are not the only types of RAID drives available. You have to look at everything available and choose based on your needs.

No one type of RAID system is going to please everybody. Though RAID is fast, cheap, and reliable, it can't be all three at once. When choosing a system, decide what your needs are. If you have budget surplus, you can afford to invest in a system that will out-perform most others. But if you're in a position where money is tight, you can still get a solid system if you choose a specific functionality.

RAID is without a doubt one of the most important technologies for server administration. Redundancy has always been a concern and now it is addressed without having to worry about an added peripheral device or time consuming, processor intensive scheduled backups. And with the rate at which data transfer speeds have increased, the high performance data writing/reading granted from data striping has helped to remove one of the biggest bottlenecks in computer networking.

More about RAID and Diffrent type of RAIDS