A closer look at RAID levels and what they mean

Anyone who's ever looked into purchasing a NAS device or server, particularly for a business, has inevitably stumbled across the term "RAID." RAID stands for Redundant Array of Inexpensive or (the more marketing-friendly term) Independent Disks.

In general, RAID uses two or more hard disk drives to improve the performance or provide some level of fault tolerance for a machine – typically a NAS or server. Fault tolerance is simply providing a safety net for failed hardware, usually a hard drive, by ensuring that the machine with the failed component can still operate. Fault tolerance lessens interruptions in productivity and the chance of data loss.

Hardware RAID

There are two ways to configure RAID – with hardware or software. There are some advantages and disadvantages with hardware-based RAID. It's more expensive, because configuring it requires an additional hardware component, a RAID controller which is a piece of hardware that controls the RAID array. RAID controllers can be internal, meaning they connect inside a server to the motherboard, or external (usually reserved for enterprise high-level RAID solutions). Hardware-based RAID is also considered a better performing, more efficient way to implement RAID than software RAID.

Software RAID

Software RAID is arguably not as reliable as hardware RAID, but it's definitely more economical and can still deliver basic fault tolerance. You can't configure RAID arrays as complex with software as you can with hardware, but if you just want to implement mirroring (which is copying data from one drive to another, to keep that data accessible in case a drive fails) then software RAID is a cheaper, less complicated to set up option.

Instead of using a bunch of disks and a controller to make an array, some software RAID solutions can use logical partitions on a single disk. That's what makes it both cheaper and less reliable – if that single disk fails completely, your data is gone.

Windows 8 and Windows 7 (Pro and Ultimate editions) have inherent support for RAID; you can set up a single disk with two partitions, and have those partitions mirrored (RAID 1) or you can setup disk striping for performance (RAID 0). This type of RAID is available in other operating systems as well, such as OS X and Linux. Since this type of RAID already comes as a feature in the OS, the price can't be beat. Software RAID can also comprise of virtual RAID solutions offered by vendors such as Dot Hill, delivering powerful host-based virtual RAID adapters. This is a solution that is more tailored to enterprise networks.

Which RAID is right for me?

Once you've decided whether software or hardware RAID best suits your purposes, you need to pick a RAID level – this refers to how you are going to configure RAID on your device. There are a number of RAID levels, and the one you choose depends on whether you are using RAID for performance or fault tolerance (or both). It also matters whether you have hardware or software RAID, because software supports fewer levels than hardware-based RAID. In the case of hardware RAID, the type of controller you have matters, too. Different controllers support different levels of RAID and also dictate the kinds of disks you can use in an array: SAS, SATA or SSD.

Here's a rundown on each level of RAID:

RAID 0 is used to boost a server's performance. It's also known as "disk striping." With RAID 0, data is written across multiple disks. This means the work that the computer is doing is handled by multiple disks rather than just one, increasing performance because multiple drives are reading and writing data, improving disk I/O. A minimum of two disks is required. Both software and hardware RAID support RAID 0, as do most controllers. The downside is that there is no fault tolerance. If one disk fails then that affects the entire array and the chances of data loss or corruption increase.

RAID 1 is a fault-tolerance configuration known as "disk mirroring." With RAID 1, data is copied seamlessly and simultaneously from one disk to another, creating a replica, or mirror. If one disk gets fried, the other can keep working. It's the simplest relatively low-cost way to implement fault tolerance. The downside is that RAID 1 causes a slight drag on performance. RAID 1 can be implemented through either software or hardware RAID. A minimum of two disks are required for RAID 1 hardware implementations. With software RAID 1, instead of two physical disks, data is mirrored between volumes on a single disk. One additional point to remember is that RAID 1 cuts total disk capacity in half: If a server with two 2TB drives is configured with RAID 1, then total storage capacity will be 2TB not 4TB.

RAID 5 is by far the most common RAID configuration for business servers and enterprise NAS devices. This RAID level provides better performance than mirroring as well as fault-tolerance. With RAID 5, data and parity (which is additional data used for recovery) are striped across three or more disks. Disk drives typically fail in sectors, rather than the entire drive dying. When RAID 5 is configured, if a portion of a disk fails, that data gets recreated from the remaining data and parity, seamlessly and automatically.

This is beneficial because RAID 5 allows many NAS and server drives to be "hot-swappable" meaning that in case a drive in the array fails, that drive can be swapped with a new drive without shutting down the server or NAS, and without having to interrupt users who may be accessing the server or NAS. It's a great solution for data redundancy, because as drives fail (and they eventually will), the data can be rebuilt to new disks as failing disks are replaced. RAID 5 can be implemented as a software or hardware solution. You'll get better performance with hardware RAID 5, because the work is done by the controller without taxing the system processor. The downside to RAID 5 is the performance hit to servers that perform a lot of write operations. For example, with RAID 5 on a server that has a database which many employees access in a workday, there could be noticeable lag.

RAID 10 is a combination of RAID 1 and 0 and is often denoted as RAID 1+0. It combines the mirroring of RAID 1 with the striping of RAID 0. It's the RAID level that gives the best performance, but it is also costly, requiring twice as many disks as other RAID levels, for a minimum of four drives. This is the ideal RAID level for highly used database servers or any server that's performing many write operations. RAID 10 can be implemented as hardware or software, but the general consensus is that many of the performance advantages are lost when you use software RAID 10.

Other RAID levels

There are other RAID levels: 2, 3, 4, 6, 7, 0+1... but they are really variants of the main RAID configurations already mentioned and used for specific instances. Here are some short descriptions of each:

RAID 2 is similar to RAID 5, but instead of disk striping using parity, striping occurs at the bit-level. RAID 2 is seldom deployed because the costs to implement it are usually prohibitive (a typical setup requires 10 disks) and it gives poor performance with some disk I/O operations.

RAID 3 is also similar to RAID 5, except this solution requires a dedicated parity drive. RAID 3 is seldom used but in the most specific types of database or processing environments that would benefit from it.

RAID 4 is similar to RAID 5 except disk striping happens at the byte level, rather than the bit-level as in RAID 2.

RAID 6 is again similar to RAID 5, but it can cope with two drive failures as opposed to just the one. However, it requires a minimum of four disks as a result (and with a four disk array, you’ll only have half the capacity usable, just the two drives), and write performance will be a tad slower.

RAID 7 is a proprietary level of RAID owned by the now-extinct Storage Computer Corporation.

RAID 0+1 is often interchanged for RAID 10 (which is RAID 1+0) but the two are not same. RAID 0+1 is a mirrored array with segments that are RAID 0 arrays. It's implemented in specific infrastructures requiring high performance but not a high level of scalability.

For most small to mid-size business purposes, RAID 0, 1, 5 and in some cases 10 suffice for good fault tolerance and or performance solutions. For most home users RAID 5 may be overkill, but software RAID 1 mirroring provides decent fault tolerance, and hardware mirroring with two physical drives provides even better fault tolerance, if you can afford it.

One last thought: Remember, RAID is not backup, nor does it replace a backup strategy – preferably an automated one. RAID can be a great way to optimise NAS and server performance, but it's only part of an overall disaster recovery solution.