What is RAID?

In this day and age, if anybody asked you what “RAID” was, you’d likely start thinking about teaming up with allies in World of Warcraft or Destiny 2, or maybe you’d start thinking about pesticide — but a few techies and hardware junkies would recognize and remember “RAID” for its acronym and its relation to digital storage.

RAID stands for “Redundant Array of Independent Disks” or “Redundant Array of Inexpensive Disks,” depending on who you ask, and they were more popular before the rise of technologies like solid-state drives (SSDs), hyperscale computing, and erasure coding. This is because RAID’s primary strength is redundant storage of the same data in different places on multiple hard disks within its array, effectively protecting against potential data loss via drive failure.

While not quite as popular as it may have been in earlier years, RAID is still utilized by governments as well as large and small businesses around the world, with major vendors like IBM, Intel, and Dell still releasing RAID products. RAID configurations afford their users the benefits of increased performance, resiliency, and affordability without sacrificing volume. According to TechTarget.com, “by putting multiple hard drives together, RAID can improve on the work of a single hard drive and, depending on how it is configured, can increase computer speed and reliability after a crash.”

What are the types of RAID?

RAID levels are used to refer to the configuration of the array, which determines how data is spread across it; either via disk mirroring (copying identical data onto more than one drive) or disk striping (partitioning each drive into smaller storage units for rapid, simultaneous access of a single file). Originally, there were only six RAID levels, spanning 0 to 5. As technology has changed, the number of levels has grown enough to warrant three RAID categories.

Standard RAID

  • RAID 0 – A configuration that offers striping across at least two disks, but no mirroring. As such, RAID 0 is used when better performance is necessary, but fault tolerance and redundancy are not.
  • RAID 1 – A configuration that offers mirroring across at least two disks, but no striping. This allows for redundancy and improved read speeds.
  • RAID 2 & 3 – These configurations both use striping across multiple disks, as well as error checking and correcting information (ECC) in some way. RAID 3 has replace RAID 2 configurations because of its use of a dedicated drive to store parity information.
  • RAID 4, 5 & 6 – These three configurations all use striping across multiple disks. RAID 4 uses larger striping, while RAID 5 uses block-level striping with parity, which means the array can still function even if one drive fails. RAID 5 requires at least three disks to operate, though five are usually recommended for optimal performance. RAID 6 uses a secondary parity scheme that allows for even further protection at the cost of performance.

Nested RAID

  • RAID 10 (1 + 0) – RAID 10 combines RAID 1 and 0, with mirrored data and striped mirrors.
  • RAID 01 (0 + 1) RAID 01 is like RAID 10, however, it utilizes striped data with mirrored stripes.
  • RAID 03 (0 + 3) RAID 03 uses RAID 0’s striping on a RAID 3 style configuration, allowing for a higher performance.
  • RAID 50 (5 + 0) RAID 05 uses RAID 0’s striping on a RAID 5 distributed parity configuration.

Nonstandard RAID

  • RAID 7 – This configuration is similar to RAID 3 and 4 with added caching capabilities. This level is owned by the now-defunct Storage Computer Corp.
  • Adaptive RAID – This configuration allows the RAID controller to decide whether to store parity in RAID 3 or RAID 5 configurations for optimal performance.
  • RAID S – Like RAID 7, RAID S is a proprietary, defunct RAID level. It’s similar to RAID 5.
  • Linux MD RAID 10 – This level allows for creation standard, nonstandard, and nested arrays for use with Linux systems.

What is RAID rebuild and recovery and who needs it?

RAID’s features allow for one hard drive to fail without destroying data or even in some cases cease operations. However, when disks unexpectedly fail, the array needs to be rebuilt with a new disk in the old disk’s place.

Even worse, when the entire array fails, the data needs to be recovered in a coherent and orderly fashion. With the sheer amount of disasters that occurred in 2017 via hurricanes, for example, you can bet that plenty experienced RAID data loss and were in need of recovery.

The data recovery specialists at Kroll Ontrack list four of the most common causes of RAID data loss:

  • Natural Disasters: As mentioned above, water, dirt, fire, and other byproducts of natural disaster can destroy a RAID like that. Recovery from this type of damage usually requires special facilities.
  • Power Issues: Loss of power, power cycling, and power surges are all major culprits of data loss. RAIDs running in a degraded state are even more at risk.
  • Mechanical Issues and Failed Rebuilds: If one disk fails or is rebuilt into the array improperly, there’s a higher chance that additional drives will also fail due to the increased workload.
  • Human Error: Last but not least, accidents, ignorance, and even malicious intent can cause an array to fail.

This is an apt time to mention that even though RAID adds protection against individual disk failure, the entire array is still vulnerable to malware, natural disasters, and sheer human stupidity. Every business, big or small, should still implement solid cyber security measures, as well as a disaster recovery plan.

Even though technology is advancing beyond RAIDs, there are still uses for it in the business world and beyond. Make sure that you’ve studied all of your options when it comes to storage before implementing any solution with widespread implications.