We return now to Storage Performance Council (SPC) results*. There have been six new SPC-1submissions, the Dot Hill® AssuredSAN™ Pro 5000, HDS HUS 150 & HUS 150 with SSDs, IBM® Power® 780 Server with SSDs, and NEC Storage M500 & M700. My apologies, it seems that I have been missing some SPC-1/E results from earlier reports which now need to be added, these two include an HP® StorageWorks™ 6400 & P6500 EVA (with SSDs). Also we have another new SPC-2 submission the SGI® InfiniteStorage™ 5600. It turns out just about every one of our top 10 charts for both SPC-1 and SPC-2 have new results.
We start our discussion with the Top 10 Least Response Time (LRT™) chart for SPC-1.
Lower is better on the Top 10 LRT chart. All of these systems are SSD only or SSD-disk hybrids. The two new systems in this top 10 include the HDS HUS 150 with SSDs and the HP P6500 EVA each of which had an all SSD backend configuration.
Over time, the LRT chart has become dominated by SSD results. It seems that even at the 10% system activity level (where SPC-1 measures LRT), the workload introduces some amount of backend overhead that is difficult to avoid. No doubt due to some amount of read miss and/or destage activity.
There was only one new result in our IOPS Top 10 and that is the IBM Power 780 Server with an all SSD backend which came in at #2 with over 780K IOPS. This is a curious storage system. The storage is attached via GX++ PCIe2 cards, which I assume are PCIe extenders to a pair of RAID controller cards in the 8-EXP30 Ultra SSD I/O drawers (which here represent the storage subsystem). In total there were 240-387GB SSDs configured in this system.
In contrast to the LRT chart above there are still a few disk only storage systems on this chart (#4-6 and 8-9). However, most of the disk only systems in the IOPS top ten had ~2000 disk drives to get to this level with the lone exception being the IBM DS8870 which had only ~1500 disks.
Next we turn to our Top 10 IOPS/$/GB
Although SPC-1 reports on $/IOPS™ we prefer our IOPS/$/GB ranking as this is less skewed by SSD/flash augmented results. In fact none of the top ten systems here have any SSD or NAND flash in them. The two new results here were the NEC Storage M700 (#4) and the HDS HUS 150 (#5) without SSDs.
This is the first time we have seen any NEC storage in SPC-1 results. The M700 had two tiers of 15Krpm disk drives (13-600GB and 691-300GB). This report was just submitted for review on 26Apr2013 so it’s very new. The only other submission from NEC was their M500 which did not reach top 10 status on any of our charts but it appears to be more of a midrange solution.
All of our primary SPC-2 Top 10 charts were changed with the new SGI InfiniteStorage submission. We start with our MBPS™ spider chart.
The new SGI submission placed #8 with a total of 8,856 MBPS. The fact that it did this with only 120-300GB 2.5” 15Krpm disk drives seems pretty significant considering it’s nearest competitors in this top 10 used 256-146GB 15Krpm disk drives.
On the other hand not sure what happened to SGI’s VOD performance that was the worst of the three by far. If they had improved this to the level of their LDQ or even LFP performance they could have moved up a number of slots. SGI doesn’t hide the fact that the InfiniteStorage 5600 is a NetApp E5560 storage system.
Finally we return to our SPC-2 MBPS/drive chart. We specifically exclude any system that uses SSDs for this ranking.
Considering SGI’s very good showing on the prior chart with a relatively modest amount of disk drives it’s no surprise that the new SGI InfiniteStorage 5600 showed up well (#1) in our top MBPS per drive chart. It appears that the IBM DCS3700 is also an E-series OEM (can’t be certain but they certainly look alike). If this is the case NetApp E-series owns the top 3 spots on our MBPS/drive ranking.
It’s good to see new SPC-1 and SPC-2 data in our top 10 charts. This new batch of SPC-1 submissions had a good showing over all for both the all-flash and disk-only systems.
Also, it’s comforting to note that HDS HUS 150 can both sustain the Top 10 low latency workload common to all SSD configurations as well as perform economically in an all-disk solution for IOPS/$/GB. This probably speaks well to the versatility of the HDS HUS 150 platform.
As always, suggestions on how to improve any of our performance analyses are welcomed. Additionally, if you are interested in more block performance details, we now provide a fuller version (top 30 results) of all these charts and a set of new ChampionsCharts™ for SAN storage in our recently updated (February 2019) SAN Storage Buying Guide available from our website. Also some of these charts are more fully discussed in our newly revised (December 2019) SAN-NAS Buying Guide also available for purchase from our website.
[This performance dispatch was originally sent out to our newsletter subscribers in April of 2013. If you would like to receive this information via email please consider signing up for our free monthly newsletter (see subscription request, above right) and we will send our current issue along with download instructions for this and other reports. Dispatches are posted to our website at least a quarter or more after they are sent to our subscribers, so if you are interested in current results you will need to signup for our newsletter.]
Silverton Consulting, Inc. is a Storage, Strategy & Systems consulting services company, based in the USA offering products and services to the data storage community.