This Storage Intelligence (StorInt™) dispatch covers Storage Performance Council (SPC) results. There have been eleven new SPC-1/1E and three new SPC-2 submissions since our last report last February. The new SPC-1/1E results are for the DataCore Parallel Server, NEC Storage M310, M310F & M710F (“F” indicates AFA), Huawei OceanStor™ 5300v3, 5500v3, 5600v3, & 5800v3, HPE 3PAR StoreServ 8440 & 8450 AFA, and NetApp EF560 AFA. All the new SPC-1 submission except for NEC M310 were all flash storage. For SPC-2, we have a new NetApp E5660 & EF560 AFA and Fujitsu ETERNUS DX8900 S3(all SSD storage).
We begin our discussion with top ten SPC-1 LRT™ (Least Response Time) results shown in Figure 1.
Figure 1 Top 10 SPC-1 LRT
The new DataCore Parallel Server is our new #1 with an LRT of 70msec and at #3 is the NetApp EF560 AFA with an LRT of 170msec. Data Core’s Parallel Server, in this submission, acted as a hyper converged, software defined storage system, with 40 SFF SATA attached SSDs (22-120GB & 18-240GB) plus a smattering of SAS attached HDDs (8-300GB). Unlikely that the disks were active during LRT testing but, 70 msec is nothing to sneeze at, even if it was all SSD IO. It’s important to note that both DataCore solutions on our LRT Top Ten (#1&2) were used in hyper converged configurations and used a software loop back block driver (assumed to be iSCSI) for IO activity that never left the combined storage-server system.
The NetApp EF560 AFA was a more traditional FC attached AFA and provided the best non-hyper-converged LRT available. It’s pretty impressive that the EF560 AFA was able to beat the IBM FlashSystem 820 with its proprietary flash module storage, but this is an older FlashSystem, the current IBM FlashSystem 900 has yet to submit an SPC-1 run.
Next, we turn to SPC-1 top ten IOPS™ in Figure 2.
In Figure 2, we can see the new DataCore Parallel Server came in at #4 and the NEC Storage M710F AFA came in at #10. There are only 3 systems in the IOPS top ten that have any disks at all, the DataCore Parallel Server (#4), the Huawei OceanStor 18800 (#6) and 6800v3 (#9). For the DataCore system there was relatively little disk storage (2.4TB disk vs. ~7.0TB of SSDs) whereas both OceanStor systems had plentiful disk storage and relatively little SSD storage. The rest of the systems in Figure 2 are all flash.
The last SPC-1 chart is the top ten $/IOPS or SPC-1 PricePerformance™ in Figure 3.
We can see a number of the new submissions in Figure 3. At #2 is DataCore Parallel Server, tied for #3, are HPE 3PAR StoreServ 8440-8450 and at #6 is NetApp EF560 AFA. Once again, we feel this metric is biased, favoring AFA’s over hybrid solutions. The DataCore Parallel Server and SANSymphony-V 10.0 are the only systems on this list with any disk at all and a relatively small amount of disk capacity as compared to SSD storage.
In Figure 4, we show our SPC-2 MBPS spider or radar chart, which depicts the achievements for the three separate workloads which make up the SPC-2 aggregate MBPS.
The new Fujitsu ETERNUS DX8900 S3 came in 1st of over 70 GB/sec in aggregated throughput made up of 84.1GB/sec of LDQ, 73.7GB/sec of VOD and 52.6GB/sec of LFP. The new DX8900 S3 is a clustered storage system with can support from 2 to 24 Control Modules. In this submission there were 4 Control Module and 192 SSDs (180-400GB & 12-800GB). It’s still unusual to see AFAs do well on SPC-2 but apparently they are getting better, because 4 of these top ten are AFAs.
In Figure 5 we can see SCI’s SPC-2 Top Ten MBPS/drive metric.
We can see the NetApp E5660, an all disk storage system, came in at #2 with 68.3 MBPS/drive, using 120-600GB 10Krpm SAS drives. There’s not much range in this metric, at least for the top ten systems, and it’s nice to see the top ten now including 10Krpm drives. We haven’t paid as much attention to drive speed before but as 15Krpm drives become less available, 10Krpm drives seem to be do just as well, for high bandwidth/throughput workloads like SPC-2
It’s always good to see new SPC-1 and -2 submissions, even if most of them are AFAs. Yet there were still some disk only and hybrid system submissions. We would have to conclude that all flash, hyper converged systems seem to have an advantage in SPC-1 LRT results as they to go out across a FC or Ethernet storage network to do an IO. Both DataCore systems in hyper-converged configurations have shown impressive LRT.
As always, suggestions on how to improve any of our performance analyses are welcomed…
[Also we offer more block storage performance information plus our OLTP, Email and Throughput ChampionsCharts™ in our recently updated (August 2017) SAN Storage Buying Guide, or for more information on protocol performance results please see our recently updated (July 2017) SAN-NAS Storage Buying Guide, both of which are available for purchase on our website.]
[This performance dispatch was originally sent out to our newsletter subscribers in May of 2016. 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 please consider signing up for our newsletter.]
Silverton Consulting, Inc., is a U.S.-based Storage, Strategy & Systems consulting firm offering products and services to the data storage community.