Fusion-io vs Intel X25-M SSD RAID, Grudge Match Review - HotHardware

Fusion-io vs Intel X25-M SSD RAID, Grudge Match Review

3 thumbs up
The IOMeter Question:
As we noted in a previous SSD round-up article, though IOMeter is clearly thought of as a well respected industry standard drive benchmark, we're not completely comfortable with it for testing SSDs, as well as comparing their performance to standard hard drives.  The fact of the matter is, though our actual results with IOMeter appear to be accurate, it is debatable whether or not certain access patterns, as they are presented to and measured on an SSD, actually provide a valid example of real world performance, at least for the average end user.  That said, we do think Iometer is a solid gauge for relative available bandwidth with a given storage solution.  Regardless, here's a sampling of our test runs with Iometer version 2006.07.27 on our SSD RAID pack versus the ioDrive.



Here we dropped in a single Intel SSD as well, for a reference baseline metric.  In our database or server access pattern, which is comprised of completely random access with 33% dedicated to write transactions, you can see the Intel X25-M RAID array scales dramatically as you add more drives to the equation and turn up the number of IO requests per target.  Even more interestingly, you can see that at a relatively low workload of 8 outstanding IOs, the ioDrive is nearly just on par with the 4-disk Intel array.  However, turn up the number of IO requests and ioDrive obliterates the Intel RAID packs with over two times the number of IOPS (Input and Output Operations Per Second).




In our Workstation access pattern, which consists of only 20% write operations and a bit more sequential access work, there are some rather interesting observations.  Again, as you add drives to the Intel RAID 0 array, performance scales relatively well, though the limitation we saw in some of our other synthetic benchmarks like HDTach and ATTO, manifests itself again, with a two drive RAID setup offering a solid performance trend even compared to the four drive array.

For the Fusion-io ioDrive, as we tax it with a larger number of requests, its performance curve just shoots for the moon.  Though we didn't plot it here, if we turned up the number of outstanding IOs to 2048, the drive actually exceeded its 100K IOPS theoretical top-end performance (103K IOPS to be exact) - which we will of course admit is just an insane amount of available bandwidth.

Article Index:

1 2 Next
0
+ -

Nice article. I really can't wait till the Fusion-io drives come down in price. I also really like that Supermicro 2.5" rack, which I think you mean fits into a 5.25" bay, not 3.5". I currently have two SSDs TAPED to the inside of my case so you can see them facing the window.

On a related note, do you think you do do a similar set of tests on different RAID controllers? I noticed you said your X58 southbridge was faster than the controller cards you had in the lab, which is odd. I really want to know the benefits of using something like the crazy Areca RAID controllers with their own upgradeable RAM sticks, like the ones DV Nation have.

0
+ -

Hey Lev,

We actually tried an Acreca 1210 card with 256MB of on board cache and the Intel array. It was actually slower than the ICH10R believe it or not. I was surprised too. However, it doesn't take much heavy lifting for RAID 0 an Intel probably has their Southbridge chipset and drivers tuned pretty well for their own SSD, so perhaps it's not all that surprising. However, with a RAID 5 setup, you definitely want hardware RAID of course.

0
+ -

And Lev, we caught that typo within like 3 seconds of go-live. You're QUICK man! LOL

0
+ -

What kind of alignment did you guys do on those Intel drives? And did you guys enable write-back cache? Those Intel drives didn't seem to scale as good as they should have after 2 were installed. Changing the alignment alone could result in some large gains. I know they are probably hitting the limits of the board, but i'm still curious if you could squeeze a little more out of them.

I didn't think that ioFusion card would be bootable. Once these kind of cards start to catch on tho... that might change with some BIOS updates, etc.

0
+ -

You guys have 4 X25-Ms. Me wants.

Anyway thats some crazy performance on the fusion io. I can't wait until stuff like that becomes affordable on the consumer end.

0
+ -

Bob, stick around for another few weeks. We'll be looking at the ioXtreme drive from Fusion-io very soon. It will be priced in the hundreds range, rather than thousands. :)

0
+ -

Dave_HH:

Bob, stick around for another few weeks. We'll be looking at the ioXtreme drive from Fusion-io very soon. It will be priced in the hundreds range, rather than thousands. :)

Thats a bit more intresting. A solid SSD would perfect my desktop I think.

0
+ -

Acarzt, write-back cache was definitely enabled and as far as alignment goes, the drives were setup with 128K stripe (default for RAID 0 on the ICH10R) and formatted with defaults for NTFS.

0
+ -

Ahh I see... That's exactly how I have mine set... it's amazing the differece write-back cache makes.

About the alignment tho... I was talking about the volume itself after creating the RAID. I don't know if this applies to the Intel drives, but I did this with my raid...

http://www.ocztechnologyforum.com/forum/showthread.php?t=53756

Vista has a default allignment of 1024. 128k is more common for RAID'd drives. I've seen guys pick up over 100MB/s by changing it. So instead of

"create partition primary align 64"

like in the walkthrough... you would use...

"create partition primary align 128"

It's worth a shot... it might make a difference :-P

0
+ -

Awesome review, $7200 wouldn't really justify for any home-use though. I saw a PCI-E add-on card where it has 8 DIMM slot for you to add old memory modules and make it into a cheap fast storage, non-bootable, and it only takes 32GB, MSRP at $499 (without the ram).

1 2 Next
Login or Register to Comment
Post a Comment
Username:   Password: