10

I am looking for some publicly available comparisons between 2.5 inch drives and 3.5 inch drives. What I had assumed is that a 2.5 inch 15k drive and a 3.5 inch 15k drive would have identical performance, however I was shown some benchmarks recently claiming that 2.5 inch 15k is considerably faster, and 10k 2.5 inch is about the same speed as 15k 3.5 inch. The problem is that these benchmarks can't be reproduced by most people because of the platform they were done on, and more to the point, it's the underlying disk that I was hoping to compare.

Basically, will a mainframe using some number of 3.5 inch 15k drives right now perform about the same as a mainframe using an identical setup of 10k 2.5 inch drives?

MDMarra
  • 100,183
  • 32
  • 195
  • 326
Basil
  • 8,811
  • 3
  • 37
  • 73
  • Just one short question: By mainframe, do you speak about some kind of System z like system, or is this just a synonym for a "regular" server? – Sven Jan 20 '12 at 15:07
  • 1
    possible duplicate of [3.5" 15k RPM drives vs 2.5" 10k RPM drives](http://serverfault.com/questions/109294/3-5-15k-rpm-drives-vs-2-5-10k-rpm-drives) – Skyhawk Jan 20 '12 at 15:31
  • The mainframe in question is a zEnterprise 196, I think. – Basil Jan 20 '12 at 15:45
  • Miles, that question didn't come up when I was entering or tagging this. In any case, drives have changed since the beginning of 2010, so the answer might be different now. – Basil Jan 20 '12 at 15:48

1 Answers1

5

Here are some benchmarks for 2.5" drives:

http://www.tomshardware.com/charts/2009-2.5-mobile-hard-drive-charts/IOMeter-2006.07.27,1130.html

Here are some benchmarks for 3.5" drives:

http://www.tomshardware.com/charts/2009-3.5-desktop-hard-drive-charts/IOMeter-2006.07.27,1034.html

Taking a look at the top 500GB drive in each section shows us that there is a difference of 36IOPS between the 104 IOPS scored by Seagate Barracuda 7200.12 and the 140 IOPS scored by the Western Digital WD Scorpio Black. The first is a 3.5" drive and the second is a 2.5" drive and they both have a 16MB cache.

Because the size of the platter is smaller, one would argue that the seek speed will be faster on a 2.5" drive, as opposed to a 3.5" drive.

It's difficult to say whether having, for example, 12 x 2.5" 10K RPM drives or 12 x 3.5" 15K RPM drives will yield the same or different performance without testing them in your environment, simulating the sort of work that will be done using these drives. Depending on the RAID configuration and RAID card used, performance could also vary wildly.

gekkz
  • 4,219
  • 2
  • 20
  • 19
  • It would be nice to have comparisons of otherwise identical drives (a WD 3.5" and a WD 2.5"), but the underlying theory is ound: Smaller platter means higher density, thus shorter seek times (all other factors being equal) – voretaq7 Jan 20 '12 at 15:56
  • I could see the contrary: longer tracks contain more data, thus as long as the file system places clusters similar data, the need for seeking altogether might be lower, even though the duration of the seek itself might be longer. From some benchmarks I've seen, seeking just one track takes miliseconds, and it's not that much smaller than seeking a full inside-to-out distance. So I'd rather have denser media, or longer tracks (or whatever allows me to do a sequential read without seeking), than lots of 'fast' seeks. I guess what I'm trying to say is there is no better seek than no seek at all. – Marcin Jan 20 '12 at 16:34
  • I'm most interested by 2.5 inch sas 10k drives (which will be available in 900GB soon) compared to 3.5 inch 15k, which are generally available in the same sizes as the 10k 2.5s. – Basil Jan 20 '12 at 16:59
  • I'm pretty sure the same results could be expected. 900GB Seagate Savvio 2.5" drives are already available in some places. – gekkz Jan 20 '12 at 21:10