Ask the Expert

RAID 10 vs RAID 50: What is the best way to configure a storage array with 16 1 TB drives?

RAID 10 vs RAID 50: What is the best way to configure a storage array with 16 1 TB drives and why? It is proposed to set up two of these boxes.

Requires Free Membership to View

RAID 10 vs RAID 50: The RAID level you use will increase or decrease the effectiveness of each of these two primary benefits -- protection and performance. The decision on RAID 10 vs RAID 50 will be driven by your requirements for:

  • How you wish to protect your data, and;
  • Consideration of the cost in disks required to deliver the RAID level protection in question.

So what are the characteristics of RAID 10 vs RAID 50?

RAID 10 (1+0): Data is striped across multiple mirrored disk sets.

In RAID 1 all data written to one disk is mirrored to its counterpart in the pair. RAID 10 takes this one step further by striping these data blocks across multiple mirrored pairs. As such, a RAID 10 set can tolerate the loss of multiple disks as long as one disk in each pair remains functional. If two disks in the same mirrored pair fail at any time, then the whole RAID 10 group is compromised and the data is lost.

RAID 10 affords excellent performance, and is considered most suitable for intensive I/O applications. The negative in using RAID 10 is that, due to mirroring, only 50% of the total raw capacity of the drives is available as usable space.

RAID 50 (5+0): Data is striped across multiple RAID 5 parity groups.

This takes a standard RAID 0 set and stripes it across multiple RAID 5 sets. As with RAID 10, multiple disk failures can be tolerated before data loss occurs as long as no more than one disk fails in any of the RAID 5 groups that make up the stripe.

This RAID type is increasing in popularity due to the benefits of increased RAID protection without the cost overheads in capacity inherent in RAID 10. RAID 50 is most suitable for fast write applications and those with a high degree of random I/O.

Whatever I/O profile you are trying to match, and whatever RAID level choice you make, it is absolutely critical that you configure at least one disk as a hot spare device to take over in the event of a failed drive. This is a basic and very effective way of ensuring that you give your data the best chance of survival when a disk fails by giving you enough time to replace the drive before another disk fails.

When considering RAID 10 vs RAID 50, the choice you make will be dependent on your I/O and protection needs. In all cases, designing an efficient storage environment is about not only knowing how much data you need to store, but understanding the type of data you are storing.

For more information: Read this expert tip comparing RAID 5 vs RAID 1.

This was first published in July 2010

 

COMMENTS powered by Disqus  //  Commenting policy