The chart shown here reflects information from a SCI StorInt(tm) dispatch on the latest Storage Performance Council benchmark performance results and depicts the top IO operations done per second per installed drive for SPC-1 and SPC-1/E submissions. This particular storage performance metric is one of the harder ones to game. For example, adding more drives to perform better does nothing for this view.
The recent SPC-1 submissions were from Huwaei Symantec’s Oceanspace S2600 and S5600, Fujitsu Eternus DX400 and DX8400 and the latest IBM DS8700 with EasyTier, SSD and SATA drives were added. Of these results, the only one to show up on this chart was the low-end Huawei Symantec S2600. It used only 48 drives and attained ~17K IOPS as measured by SPC-1.
Other changes to this chart included the addition of Xiotech’s Emprise 5000 SPC-1/E runs with both 146GB and 600GB drives. We added the SPC-1/E results because they execute the exact same set of tests and generate the same performance summaries.
It’s very surprising to see the first use of 600GB drives in an SPC-1/E benchmark to show up well here and the very respectable #2 result from their 146GB drive version indicates excellent drive performance yields. The only other non-146GB drive result was for the Fujitsu DX80 which used 300GB drives.
Also as readers of our storage performance dispatches may recall the Sun (now Oracle) J4400 array provided no RAID support for their benchmark run. We view this as an unusable configuration and although it’s advantages vis a vis IOPS/drive are probably debatable.
A couple of other caveats to this comparison,
- We do not include pure SSD configurations as they would easily dominate this metric.
- We do not include benchmarks that use 73GB drives as they would offer a slight advantage and such small drives are difficult to purchase nowadays.
We are somewhat in a quandary about showing mixed drive (capacity) configurations. In fact an earlier version of this chart without the two Xiotech SPC-1/E results showed the IBM DS8700 EasyTier configuration with SSDs and rotating SATA disks. In that version the DS8700 came in at a rough tie with the then 7th place Fujitsu’s ETERNUS2000 subsystem. For the time being, we have decided not to include mixed drive configurations in this comparison but would welcome any feedback on this decision.
As always, we appreciate any comments on our performance analysis. Also if you are interested in receiving your own free copy of our newsletter with the full SPC performance report in it please subscribe to our newsletter. The full report will be made available on the dispatches section of our website in a couple of weeks.
Seriously, you don't include all-SSD configs because they'd always win? And you don't include mixed-drive configs because….well, just because?
____
Yet you allow 600GB 15K rpm drives that are not fully utilized…go figure.
____
Put 73GB on a 73GB 15K rpm drive vs. the same 73GB on a 600GB drive, and which will perform faster? So, you'll allow short-stroking drives, but not using SSDs. I suspect this goes back to the roots of the SPC-1, when it was optimized to ensure that the working set would exceed the projected capacities of large-cached disk arrays. SSDs allow more systems to utilize (effectively) larger caches, and so you exclude them.
____
And this is helping customers to choose the appropriate storage exactly how?
As for SSDs there really is no comparison between these two technologies. And yes short stroking can game this metric but it hurts in the other popular metric of IOPs/$. But I only discuss one chart from my analysis. Thanks again for your comment.
Ray Lucchesi
720-221-7270
Sent from iPhone
@storageanarchy, if you are talking about the 600GB Xiotech drives, your information is wrong. The SPC-1/E tests from Xiotech was done at 98% capacity, i.e. no short-stroking, and it was a 10k drive. Yet still great performance, hmm.
Mickey, Thanks for the comment. But what wrong information do your mean?
Ray Lucchesi
720-221-7270
Sent from iPhone
Mickey – I wasn’t picking on anybody’s results.
I was merely making the point that selectively excluding configurations isn’t necessarily serving the customers’ interests.
And Ray, if you’re going to use $/IOPS to neutralize the benefits of short-stroking in the comparisons, then why not also use $/IOPS to put SSD vs. HDD on a level ground?
Perhaps the more interesting question would be whether the system that gets the most IOPS and lowest response times per HDD spindle is also the one that gets the most IOPS and lowest response times per SSD. I’ll bet One US Dollar that there is little correlation between the two – at least for the majority of storage systems shipping today.
I think Mickey was pointing out a flaw in @StorageAnarchy's comment around the 600GB solution not being fully utilized. The Datapac that was tested was a 600GB solution running at 98% utilization. ie) no short stroking/cheating etc. Also, he assumed it was 600GB 15K drives due to the nice IOPS results but it was actually our 10k RPM drives which makes this even more impressive. Then again, i'm a little biased – i work for Xiotech 🙂
Nice write up !!
@StorageTexan
Tommyt/StorageTexan, Yeah, I realized after I responded to his comment that he might have been talking about the comment and not the post. Thanks for pointing that out.
Ray
Sent from my iPad.
Ray,
@storageanarchy was implying that the 600gb drives that Xiotech used for the SPC result were short stroked in their SPC test results. Since the SPC results are published publicly and can be reviewed for the configuration, Mickey simply fact checked the SPC result that Xiotech submitted and found that the drives were FULL STROKED at 98% capacity!!
Frankly this is an extremely impressive feat I have never heard of a result over 80% before and most are as low as 50% or less even…ie SHORT STROKED. So to getthis kind of performance on a fully utilized disk drive is a tremdous feat. Good for them.
However I would agree with @storageanarchy on one point. SSD and mixed tier configuraitions should be included. I think there is an assumption that they will blow everyone out of the water but isn't that what we are looking for?? If they want to be fair than setup seperate catagories but Enquiring Minds Want to Know!!
Gunslinger,
I guess we will have to disagree on including SSDs in this metric. The fact that SSDs can offer almost 100X the IOPS/drive that normal drives can sustain argues to put them in a separate category. There are plenty of other charts in my longer reports which includes SSDs and mixed drive configurations.
Ray
Sent from my iPad.
Hi there
I am wondering something
Here you debate about EMC, NetApp, IBM, HP, and I do not see FUjitsu nowhere
Is it really that bad?
I am wondering because we might be buying one of their storage products (DX8400 Eternus), and I am wondering how are experiences with this kind of storage. I haven't found much on the internet, other then data-sheets, and I have seen that they haven't been accepted on SPC yet.
So, the big question is – does anyone know, or has any experience with this sort of enterprise storage?
Thanks
Alex
Alex,
On my blog I only publish one performance chart a month. In my monthly free newsletter I have much more information, if you want to receive a copy just subscribe by sending me an email. As for the DX8400 it has completed and published an SPC-1 benchmark (A00093) which is freely available online at the SPC website noted in the post above. I suggest you look to that if you want to see how it did. As for the IOPS/Drv results that I report on in this post, all I can say is it did not come in the top 10.
Remember there are a number of dimensions or metrics that can be used to measure storage performance. It helps to know what your workload is in order to understand which ones may be most important to your data center. But, no benchmark is going exactly duplicate your environment. no matter what. However, benchmarks like SPC, SPECsfs and others are the only independently verified/audited performance runs we as analysts can see and so are the only way for us to compare performance across subsystems.
Ray