We return now to the preeminent block storage benchmark, Storage Performance Council (SPC) results*.  There have been four new SPC-1submissions, the Huawei Symantec OceanSpace™ S5600T, Huawei Symantec OceanSpace S6800T, Huawei OceanStor™ S6800T, and IBM DS8870, plus two new SPC-2 submissions, the SGI Infinite Store 5500 and the IBM DS8870.  As at least one of these seemed to show up in most of the top 10 chart we review so we cover both SPC-1 and SPC-2 results for this report.

Unclear what the difference is between the Huawei Symantec OceanSpace S6800T and the Huawei OceanStor S6800T but this could be just rebranding. However, the Huawei OceanStor submission used more drive spindles (768 vs. 368) than the OceanSpace submission.

SPC-1*results

We start our discussion with the top 10 I/O operations per second (IOPS™) performance for SPC-1.

(SCISPC121029-001) (c) 2012 Silverton Consulting, Inc., All Rights Reserved

Figure 1 SPC-1* Top 10 IOPS

Higher is better on the IOPS chart. The only new showing here was the IBM DS8870 which came in at #4 with ~451K IOPS.  All the rest of the entries on this Top 10 chart have been reviewed before, in prior dispatches.  Recall that the K2-D is a DRAM storage system.

There was no change to the LRT charts, which these days are dominated by flashcache or SSD systems.  So, we turn our review to IOPS/drive results

(SCISPC121029-002) (c) 2012 Silverton Consulting, Inc., All Rights Reserved

Figure 2 SPC-1 Top 10 IOPS/drive results

Here we can see all the latest Huawei submissions at #1, 2 and 9.  Recall that the IOPS/drive only reports on submissions with >140GB drives and no SSDs/FlashCache in their configuration.

All the Huawei systems were configured with two different drive capacities at 300GB and 600GB drives.  Why different drive capacities would help them do well here is an open question. Also not sure why the latest OceanStor S6800T system (@#9)seemed to do so poorly as compared to its brethren (@#1&2) but it did have more drives and did supply higher IOPS (although not top 10 level).

Next we turn to storage IOPS/$/GB.

(SCISPC121029-003) (c) 2012 Silverton Consulting, Inc., All Rights Reserved

Figure 3 Top 10 SPC-1 IOPS/$/GB

In contrast to the SPC-1 reported $/IOPS metric, we prefer our own IOPS/$/GB.  For one thing it’s not as biased to flash arrays.  In any event, all three new Huawei submissions placed well here at #3, 5 and 6.  With previous Huawei submissions, SPC-1 report pricing was supplied in Chinese Yuan (CNY) and were converted to USD$ at the time of analysis.  But all new Huawei submissions were quoted in USD$ so that’s no longer a concern.  It’s interesting to see that five out of the top 10 IOPS/$/GB are from Huawei, probably deserves further investigation.

SPC-2 Results

We turn now to the SPC-2 throughput results.

(SCISPC121029-004) (c) 2012 Silverton Consulting, Inc., All Rights Reserved

Figure 4 Top 10 SPC-2 MBPS™

The new IBM DS8870 came in as the new #1 with over 15GB/sec of data throughput.  One surprise is that the DS8870 used 146GB disk drives, didn’t know they were still around.  The HDS VSP systems (includes the HP P9500 XP Array) and the Oracle ZFS storage both used 300GB SAS drives with the IBM XIV had 2TB SATA drives.  Of course the now IBM TMS RamSan-630 was an all SSD system.  IBM has four different systems in the SPC-2 Top 10 MBPS with its SVC running Storwize V7000s, the two IBM DS8000 series (DS8870 and DS8800) and the IBM XIV Gen 3.

Another metric we occasionally like to see is the MBPS per drive chart.

(SCISPC121029-005) (c) 2012 Silverton Consulting, Inc., All Rights Reserved

Figure 5 Top 10 SPC-2 MBPS/drive

Similar to the discussion in the SPC-1 section above, this chart plots “disk-only” submissions, with no flash cache or SSD  and over 140GB disk drives.  The new SGI InfiniteStorage seemed to do well here, generating almost 68 MBPS/drive with their #1 result.

Significance

It’s good to see new SPC-1 and SPC-2 data in our top 10 charts.  Also it’s unusual to have so many “disk-only” systems perform so well.  This seems to indicate there’s still a place for disk-only storage in enterprise environments.

All-flash arrays seem to be falling off these charts, but we have yet to see submissions from most of the new SSD startups, except for Kaminario, let alone some of the older SSD players like IBM’s TMS acquisition. When the startups  benchmark their systems competition should heat up.

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 new ChampionsChart™ for SAN storage in our recently updated (August, 2014) SAN Storage Buying Guide available from our website[1].

~~~~

[This performance dispatch was originally sent out to our newsletter subscribers in August of 2012.  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.  

Silverton Consulting, Inc. is a Storage, Strategy & Systems consulting services company, based in the USA offering products and services to the data storage community.