How RAID Storage Protects Data: Understanding Benefits and Trade-Offs of RAID for Data Protection

There seems to be quite a lot of hype about the benefits of RAID storage solutions, and to be fair, within the context of business-class server systems, RAID storage has provided huge benefits. The critical part is understanding the trade-offs to avoid unmet expectations, and this is even more important before laying out the money for RAID storage on a single-user PC.

RAID Benefits in a Nutshell

The main goals of RAID storage configurations are to reduce downtime or lost data due to hardware failure, and to improve performance. It is definitely true that the right level of "parity" can prevent loss of data due to a disk failure. It is also true that the striping features of some RAID levels can improve response time or throughput for some application

However, this hardware protection and performance gain isn't foolproof, nor is it a given for many end user systems. For example, it is important to realize that RAID itself does not "backup" data, so any data corruption or accidental deletion will still cause lost data - meaning that backup strategies are still essential.

RAID Protection from Data Loss

There are several RAID Levels defined, however generally only RAID-0, RAID-1, RAID-1+0, and RAID-5 are used. There is also RAID-6, which is really just adding an additional parity block to a RAID-5 array to allow for more than one simultaneous hardware failure. This has some additional trade-off of course.

RAID-0 does not protect against any data loss whatsoever, so it should not be considered if that is the goal for implementing a RAID solution. Statistically speaking, the more disk drives involved in an environment, the more likely there will be a failure, and a RAID-0 configuration is completely lost if even one of those drives fails.

RAID-1 provides excellent protection from data loss due to hardware failure since the statistical probability is very low that both drives involved in that mirror will fail simultaneously. Everything written to the RAID-1 "drive" is actually written to both drives nearly instantaneously - and this means any data corruption or accidental file deletion (or virus corruption).

There is also RAID-1+0, which is where multiple RAID-1 sets (two drives each) are striped into a RAID-0 set to provide greater performance benefit. While it is true that increasing the number of drives in the RAID-1+0 set increases the probability of multiple simultaneous failures in the whole array, the probability of both drives in a single RAID-1 set failing simultaneously is very small, allowing this configuration to withstand multiple simultaneous failures.

RAID-5 (or RAID-6) will give you similar protection in that there will be additional parity of the data interspersed throughout all drives in the RAIDset. The minimum number of drives needed is 3 (or 4), otherwise it would really just be a RAID-1 set. The more drives in the RAIDset, the greater the probability of multiple simultaneous drive failures, and a RAID-5 (or RAID-6) array can only withstand a single drive (or two drive) failure without data loss.

RAID Protection does not Eliminate Backups

It is critical to realize that having a RAID storage system of any variety will not eliminate the need to perform backups. While the data blocks on disk are protected by additional parity copies, these parity copies are updated when any update occurs on the original data block - i.e., when a file is corrupted, infected with a virus, or deleted. A RAID system will happily copy all of this activity to the parity block.

Other RAID Considerations

Aside from the data protection benefits of RAID storage solution, there are also performance and cost considerations. These two will be explored in later articles in order to provide the proper level of detail that each requires.