Disk array dimensions. Narrow vs wide and tall vs short arrays.

You have probably heard the unscientific descriptions of storage arrays of various types, describing arrays as "tall" or "wide" or some such. These descriptions come from the way we commonly draw array layout diagrams on paper. Disks are drawn as vertical rectangles of fixed width. The larger the disk, the taller the rectangle.

A0 A1 A2 A3 A4 A5
B0 B1 B2

The example shows a sample drawing of two arrays side by side.
The left one uses many small disks, and the right one uses a few big disks.

The array is said to be

  • narrow if there are few disks in it;
  • wide if there are many disks in it;
  • short if individual disks are small;
  • tall if the individual disks are large.

There is no specific boundary where short becomes tall and vice versa, or where narrow becomes wide.

However, these terms come in handy when the array geometry is skewed enough to become a factor by itself. Consider the following examples:

  • The mirror of two 16 TB drives is tall and narrow. There is nothing inherently wrong with it, but data recovery will take a long time to scan the entire 16TB, and there is no way to parallelize the process.
  • The RAIDZ2 vdev (or RAID6 array) of 16x 1TB drives is short and wide. It is probably a little wasteful on physical space and power, but the performance should be good.

Created 3 October, 2021

I have a low volume mailing list, for news and tips, which I send out once or twice a month.
Subscribe if you are interested.