SPEC SFS®2014 Performance Report as of June 2020

In ORT, SPECsfs2014, Streams, Streams MB/Sec, VDA workload by AdministratorLeave a Comment

This Storage Intelligence (StorInt™) dispatch covers SPEC sfs2014 benchmarks . There was one new vendor submission for the VDA (video data acquisition) workload, the DATACOM INFINITY storage system.

SPEC sfs2014_vda benchmark results

The SPEC sfs2014_vda workload simulates video data streams. The VDA workload consists of two activities, one is high-def, camera data streams (100% write) and the other is companion applications (~84% random read). We could find no specification for the ratio of data streams to application IO, but we believe it’s at least 80% stream IO.

Figure 1 shows the top 10 systems for VDA maximum concurrent stream supported, where each stream equates to a single camera being recorded.

Figure 1 SPECsfs2014_vdai top 10 maximum number of data streams supported

In Figure 1, the new DATACOM INFINITY unified storage system came in as our new #3 with 2800 cameras streams supported. The INFINITY storage system is hybrid, clustered storage system with both SSDs (NVMe) and disk storage.

In their benchmarked configuration, there were 10 storage nodes. Each node had one 800GB NVMe SSD for storage logs (journals), 2 240GB SAS SSDs for OS software, and 36 4TB 7200RPM disk drives for data storage using mirrored data protection, for a total of 360 disks for data storage across the cluster. The host connections were via 10GbE, 4 per node. The storage cluster supported 625TiB of usable capacity. Each storage server node had 256GiB of system memory for a total storage cluster capacity of 2.6TiB of memory.

In comparison, the DDN SFA14KXE (#2 in the chart above) was a disk-only system but it had more disks (454) and much more capacity (3600TiB) than the DATACOM storage. The WekaIO (#1 above ) was an all NVMe SSD, clustered storage system.

Next, we show the top 10 Overall Response Time (ORT) for VDA submissions in Figure 2.

Figure 2 SPEC sfs2014_vda top 10 VDA workload overall response times (ORT)

In Figure 2, the DATACOM system came in at #6. Recall that ORT is an average across the whole benchmark series (from lowest to max stream counts) or at ~50% of the max workload. The DATACOM 16.7msec response time was generated for ~1680 camera streams.

The #2, SPEC SFS subcommittee solution maxed out at 100 camera streams so it’s not a great comparison. Neither is the #1 WekaIO solution as it was an all NVMe only storage. The two IBM Spectrum solutions (#3 & #4) were both AFA storage, so also not fair comparisons. On the other hand, the Cisco MapR-XD (#5) (maxed out at ~2K streams) was disk only storage, so compared to that one, the DATACOM did ok. The DDN system didn’t reach the top 10 in ORT.

In Figure 3, we show the top 10 systems in max MB/Sec throughput for the VDA workload

Figure 3 SPEC sfs2014_vda top 10 VDA MB/Sec systems

In Figure 3, we also show both the minimum and maximum system throughput. The DATACOM storage came in at #3, right behind DDN SFA14XXE storage. This chart mimics the rankings of Figure 1 above, which implies that each stream is generated ~4.6MB of data per second.

We believe VDA workloads are mostly data write throughput intensive, which should favor disk storage over all AFA systems, whose SSDs have problems with heavy write workloads. The fact that the WekaIO, an all NVMe SSD storage did so well here, says something about how they have architected their storage system to support write activity.
Significance
It’s always great to see new SPEC sfs2014submissions, even if there are only one of them. It gives us a chance to see and compare a set of storage systems under a defined load

We can’t seem to find much information about DATACOM’s storage product line. There are a number of companies throughout the world using the DATACOM name. None of them that we found had an INFINITY storage product line. It would be helpful if SPEC SFS had a link to the company’s website or product website in its reports.

We are still trying to determine the best way to report SPEC sfs2014 results and as such, we may experiment with variants of the above charts. Any ideas on other metrics of interest to report, please do let us know. Furthermore, suggestions on how to improve any of our performance analyses are always welcomed.

This report was sent out to subscribers as part our free, monthly Storage Intelligence e-newsletter. If you are interested in receiving future storage performance analyses along with recent product announcement summaries, please use the QR code (below right) to sign up for your own copy.

[This storage performance was originally sent out to our newsletter subscribers in June of 2020.  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. ]  

Silverton Consulting, Inc., is a U.S.-based Storage, Strategy & Systems consulting firm offering products and services to the data storage community

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.