RAID – What is it?
Many people have heard of RAID drives before, but there is a huge amount of confusion out there as to what it is and what it does, and it is really very important to understand. For extensive technical information on the different RAID levels and how they work have a read of the Wikipedia RAID pages.
RAID (Redundant Array of Independent Disks) – What is it?
RAID is a method of using multiple disks together to achieve redundancy i.e. preventing data loss if one disks fails. RAID “levels” are different ways of joining those disks together, and each has different characteristics in terms of data-security, write-speed, read-speed, cost, etc.
Some RAID levels can be enacted in software (usually RAID-0 and RAID-1) and whilst this is obviously cost-effective, a proper hardware RAID controller provides a more reliable solution.
For software-RAID under MacOSX you can create RAID-0, RAID-1 & JBOD volumes directly from the Disk Utility, or use third-party software such as SoftRaid.
Here is a summary of the most common RAID levels: RAID-0, RAID-1, and RAID-5, there’s also a good explanation here:
A typical RAID-0 setup might use 2 x 250GB drives “striped” into a single 500GB RAID-0 volume. In this setup when you save a 100MB file to the volume, it splits it into two 50MB segments and writes one segment to one drive, and the other segment to the other drive.
Advantages: Because you can read and write data to multiple drives simultaneously, it is very fast – with two drives it will usually be 90% faster than the speed of a single drive. This is great for video and other work requiring huge amounts of data to be moved around. Also, you keep the full capacity of both drives (500GB in this example).
Disadvantages: If EITHER drive fails you have lost ALL of the data on BOTH drives. You do **NOT** want to use RAID-0 for any sort of image archiving or other backup purposes.
A typical RAID-1 setup might use 2 x 250GB drives “mirrored” into a single 250GB RAID-1 volume. In this setup when you save a 100MB file to the volume, it writes that full 100MB file to both disks simultaneously, thus at any time each disk is a perfect mirror of the other.
Advantages: If one drive fails, a full copy of the data is still on the other drive, so you haven’t lost anything at all. Just replace the dead drive, and the system will auto-rebuild your two drive mirror so that you are “protected” again. Faster read speeds than a single drive, but slower writes.
Disadvantages: You “lose” half the capacity of your drives because with 2×250GB drives, you only get 250GB of RAID-1 storage.
A typical RAID-5 setup might use 4 x 250GB drives formatted as a 750Gb RAID-5 volume. In this setup when you save a 100MB file to the volume, it spreads it out over all four disks, but also writes special “parity” data across all four drives, so that if any single drive fails the system can rebuild all the lost data from the dead drive.
Advantages: Any single drive can fail and you don’t lose anything. In terms of storage space, you only “lose” the capacity of one drive, so that with 4×250GB drives, you get 750GB of useable storage, and thus is more space-efficient than RAID-1.
Disadvantages: Hardware is more expensive than RAID-0 or RAID-1 because it requires a more powerful controller to compute the “parity” data every time data is written to the volume. On the other hand it’s more cost-effective in terms of storage capacity. Requires a minimum of three, usually four, hard drives and cannot match the speed of RAID-0 or 1.
Other less common RAID levels
JBOD: “Just a Bunch Of Disks” and also called Concatenation. This is not really RAID at all, but is a popular method for combining multiple physical disk drives into a single virtual volume. As the name implies, disks are merely joined together, end to beginning, so they appear to be a single large disk and one can think of it as the opposite of partitioning. It provides no redundancy if drives fail.
RAID-2: Pretty much obsolete.
RAID-3: Similar idea as Raid-5 but poorer performance, rarely used.
RAID-4: Like RAID-5 except the parity data is stored on one drive not spread across all. Generally considered to have been superseded by RAID-5 because of the latter’s better performance.
RAID-6: Like RAID-5 except that even more parity data is stored so that up to two drives can fail without any data loss.
RAID-0+1, 1+0, 10, 50 etc: These are “Nested Raid Levels” that combine one or more RAID types and usually involve a large number of disks.
unRAID: Not a RAID level, but a proprietary implementation by Lime Technology similar to RAID-4 except that the data is not striped across the disks. Think JBOD with parity protection. Quite interesting actually, but few products are available.
MatrixRAID: Not a RAID level and a technology possible in theory by many RAID controllers but popularised by Intel who incorporated it into some of its chipsets, combining a RAID-1 volume and RAID-0 volume on the same set of two or more disks.
RAID is really useful, but an important point to remember about ALL the RAID levels is that they are designed to prevent data loss through drive failure. They will NOT prevent you losing your work via any of the corruption/human error examples listed on page 1.
For this you need a manual or automatic regular backup strategy, so that you can go back to how your files were before you made the error.
Next page: 3. How to create a backup strategy
Return to “RAID for Photographers” index page
