I previously posted about various approaches of protecting your data from loss. The issue as I mentioned is a risk of hardware failure, user error or virus. All these are major issues, but the second two can be handled by carefully what is done on the computer. The first problem, hardware failure is more random in nature and there is little you can do to prevent it. RAID is a method to protect you from hardware failure.

RAID stands for “Redundant Array of Independant Disks” and is a family of algorithms designed to leverage multiple disk drives. Each algorithm is called a “RAID level”. Here is a summary of the RAID levels:

  • RAID 0 – This is the most basic form of RAID and does not provide any redundancy and instead is focused on performance. (Assume a RAID 0 configuration containing 2x250GB drives) RAID 0 takes the drive resources and pools them into one large volume. Thus in this case, the computer would see one large 500 GB drive. As data is written to the drive, it will be written in a round robin fashion to each drive in the pool. This provides for better performance since you are writing to two drives instead of one, but also provides greater risk since each drive has half your data and if you lose one drive, you will corrupt all your data.
  • RAID 1 – This approach provides a high level of reliability for data at the cost of more disk space. (Assume a RAID 1 configuration containing 2x250GB drives)RAID 1 mirrors data between two drives. Thus as data is written to the first drive, the same data is simultaneously written to the second drive. The result is that the two drives are exactly the same and either one cane be used to access the full data set. The downside is that capacity is reduced in half compared to RAID 0, and in this example, you would see 250 GB of available storage. This is the approach I use.
  • RAID 5 – This is a more complex approach than the other two. It uses a complicated algorithm to create mathematical data that can be used to re-construct data in the case of a failure. It is designed to augemt RAID1 by providing better capacity. Because of the complexity in calculating these mathematical equations, RAID 5 typically requires dedicated hardware to perform the calculations. This is the most common RAID technology you see in corporate environment. Another side note is that this approach typically requires at least 3 hard drives.

Should I use RAID?

Yes, I think that RAID is a very important technology and encourage all readers to consider it.

Which RAID level should I choose?

There is no question that RAID1 is the best choice for most users. In larger environments, RAID5 would typically make more sense.

How do I enable RAID?

There are different ways to enable RAID. One method to create a RAID set is to use software. Most OSes include the ability to create a RAID set in software. This approach should only be considered for RAID0 or RAID1, and is generally the least reliable approach and should be avoided in a Windows environment. A number of vendors make dedicated RAID cards that you plug into your computer and will handle the creation and management of RAID sets. These are inexpensive and are recommended. The third option is that some motherboards included embedded RAID hardware which makes the creation and management the easiest. My recommendation is to either purchase a RAID card or use the embedded feature in your motherboard, if available.

(Visited 49 times, 1 visits today)