English
+370 5 205 5502 sales@monovm.com

In this article we will go in-depth on what is RAID technology, how it is used and its benefits and downsides. We will also mention the most used RAID configurations and the differences between them.

15

Apr, 19

What is RAID and Why Use It

  • by Antoniy Yushkevych
  • in Server
  • View 638

It is not a secret that technology has been advancing at a rate previously unseen. All the performance numbers of all the hardware components we use today have risen annually in an exponential manner. Nevertheless, just like how in the automotive industry, packing a vehicle with pure horsepower won’t win you the race, in the IT world, performance improvements are not simply due to better hardware. There are countless incredibly genius solutions that can cause even old outdated hardware to outperform the “latest and greatest” on the market. Today we will focus on one of such solutions: Redundant Array of Independent Disks (i.e. RAID) Technology.

 

What is RAID?

RAID stands for Redundant Array of Independent Disks and is a technology originally created to provide a level of fault tolerance and improved performance for server machines. Fault tolerance simply means that if a hard disk was to fail, the machine would still continue operation.  The level of fault tolerance and performance benefits you receive from your RAID array depends on how it is set up and whether you are using hardware RAID controllers or software RAID solutions.

 

RAID works by connecting two or more (the minimum amount depends on the RAID configuration used) hard drives and striping, mirroring them or both. Striping is the process of writing data across multiple disks, meaning the disk I/O is significantly improved. Disk mirroring, on the other hand, causes the data to be copied simultaneously from one disk to another, creating a mirror. Different RAID configurations utilize these two technologies in different ways to create custom ratios between data redundancy and performance improvements.

 

Hardware vs Software RAID

Software RAID allows you to set up a RAID array without needing to use a dedicated hardware RAID controller, meanwhile providing similar functionality. It comes at a price though; the performance benefits gained from software RAID are not nearly as great as the ones you could get when using a specialized hardware solution. Another major issue with software RAID is the fact that it is software, therefore there will be no data redundancy during boot time, as the software has not launched yet. Despite that, RAID capability is inherent in the operating system. Windows has had built-in support for RAID since Windows 7 (Pro and Ultimate editions). It allows you to split a single disk into two partitions: one to boot from and another for data storage. Then you can have the data partition mirrored.

Hardware RAID typically uses a discreet hardware controller, however ROC (i.e. RAID-on-Chip) are also widely used in the IT world.  In ROC solutions, the RAID processor, memory controller, host interface, I/O interfaces for hard disk drive connections, and sometimes even the memory are all integrated into one single chip, which can be integrated into the motherboard and offers hardware RAID capabilities.

 

 

Types of RAID Arrays

These are the most used RAID configurations.

Features

RAID 0

RAID 1

RAID 5

RAID 6

RAID 10

Minimum # of drives

2

2

3

4

4

Data Protection

No Protection

Single-drive failure

Single-drive failure

Two-drive failure

Up to one disk failure in each sub-array

Read Performance

High

High

High

High

High

Write Performance

High

Medium

Low

Low

Medium

Capacity Utilization

100%

50%

67%-94%

50%-88%

50%

 

Different RAID configurations in detail:

RAID 0: This RAID array is used purely to increase server performance and does not provide any data redundancy, thus providing 100% of capacity utilization.

 RAID0

RAID 1: This RAID array does not do any striping but uses the second drive purely for data mirroring.

 RAID1

RAID 5: Data parity is striped across three or more disks, so if one has an error, data is recreated from this distributed data and parity block. This means that the drives in this array are hot swappable.

 RAID5

RAID 6: Similar to RAID 5, but with one more parity block, meaning that 2 drives can fail and the server will still be operational.

 RAID6

RAID 10 (1+0): It is a combination on RAID 1 and RAID 0, often called RAID 1+0. This is the most optimal solution as it combines data striping and data mirroring to provide the best performance and data redundancy. Unlike the rest of RAID arrays, it requires a minimum of 4 hard disks.

 RAID10