Hands on: Protecting your data with Raid

Our recent articles benchmarking various Raid configurations sparked a debate in the Hands On mailbox.

Long-term Raid users defended the technique of using multiple hard disks as the only certain way to protect against disk failure.

Others believed Raid was overkill for general users and argued that copying files onto an external hard disk was a much better use of resources.

As with all good arguments, there are cases for both sides. Raid can be perfect for some situations, but overkill for others. The important thing is to have some kind of solution in place to protect your data against disk failure. So, with this in mind, we’re returning to the subject of Raid, but this time weighing up its pros and cons against simpler solutions to see if it’s the right technology for you.

Basics of Raid
In previous issues, we’ve described in detail how Raid works, so we’ll just cover the basics here. Raid stands for Redundant Array of Inexpensive or Independent Hard Disks and, in fact, uses two or more disks to deliver improved performance, reliability or both.

All Raid arrays are managed by a Raid controller, which does the work of distributing the data between the disks and allowing them to be treated by an operating system as a single storage volume. Raid controllers come in many shapes and sizes, and most motherboards now feature one of some description as standard. Alternatively, you can equip a PC with Raid facilities by using an expansion card from companies such as Promise.

Budget Raid controllers provide little more than a number of plugs to connect the hard disks and offload the data management to software running under your PC’s main processor. More expensive controllers feature dedicated hardware to handle the disk and data management to deliver much improved performance. Different Raid controllers also support different types of Raid arrays.

The most common type of Raid arrays are known as Raid 0, Raid 1 and Raid 5. Raid 0 reads and writes data on all disks simultaneously, improving speed and making the full capacity available, but at the cost of reliability – if one disk fails, you lose the whole array. Raid 1 makes an identical copy of one disk on another to provide 100 per cent redundancy, albeit at the cost of half the total capacity and a slower speed than Raid 0.

Raid 5 uses three or more disks and delivers 100 per cent redundancy by writing parity data on all of them. The advantage over Raid 1 is losing only one disk’s worth of capacity to redundancy, but the calculation of parity data is complex and delivers poor write performance under software-based controllers. Hardware-accelerated Raid 5 controllers solve the performance issues and can deliver read and write speeds approaching Raid 0, but decent models can cost more than £200.

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!