STAC Report: kdb+ 4.0 on 8 x Dell PowerEdge R7525 servers accessing a 3-node Dell PowerScale F900 All-Flash (tick analytics)

Dell PowerScale F900 at scale.

13 September 2022

As a follow-on to last year’s baseline STAC-M3™ benchmarks on a PowerScale F900-based solution, Dell Technologies asked STAC to perform the scaling suite to demonstrate the F900 with larger datasets. This time, the database was on 8 Dell PowerEdge R7525 servers with AMD Epyc™ 7H12 CPUs. The results are available here.

STAC-M3 is the set of industry standard enterprise tick-analytics benchmarks for database software/ hardware stacks that manage large time series of market data ("tick data"). In this project, we ran both the baseline benchmark suite (code named Antuco) and the optional scaling suite (code named Kanaga).

The stack under test (SUT) was KX’s kdb+ 4.0 DBMS distributed across 8 Dell PowerEdge R7525 servers, each with 2 x 64-core AMD EPYC™ 7H12 CPUs, sharing access to a Dell PowerScale F900 All-Flash Scale-Out NAS 3-node cluster with 251TiB total physical storage capacity. 100 GbE was used for communication, and the database nodes accessed the data over standard NFS. The SUT used Compatibility Revision H of the kdb+ STAC Packs. Dell Technologies chose to highlight that this solution:

• Was faster than an on-prem solution involving 4 database servers accessing an earlier generation of Dell Technologies’ flash storage appliance with kdb+ 3.6 (SUT ID KDB190430) in 14 of 17 mean-response time STAC-M3 Antuco benchmarks, including:
• 16.9x speed-up in the 10-user market snapshot (STAC-M3.β1.10T.MKTSNAP.TIME)
• 4.1x speed-up in the 100-user unpredictable interval stats (STAC-M3.ß1.100T.STATS-UI.TIME)
• 3.4x speed-up in the 50-user unpredictable interval stats (STAC-M3.ß1.50T.STATS-UI.TIME)
• When compared to a cloud-based solution involving kdb+ 4.0 in sharded mode on 12 database servers accessing data on local NVMe SSDs (SUT ID KDB211210):
• Was faster in 4 of 17 mean-response time STAC-M3 Antuco benchmarks, including:
• 2.9x speed-up in NBBO (STAC-M3.β1.1T.NBBO.TIME)
• 2.4x speed-up in 1-user unpredictable interval stats (STAC-M3.ß1.1T.STATS-UI.TIME)
• Was faster in all 5 mean-response time STAC-M3 Kanaga benchmarks involving 100 users
  (STAC-M3.β1.100T.YR{1,2,3,4,5}VWAB-12D-HO.TIME)
• Was faster in 3 of 5 mean-response time STAC-M3 Kanaga benchmarks involving 50 users
  (STAC-M3.β1.50T.YR{3,4,5}VWAB-12D-HO.TIME)

For details, please see the report at the link above. Premium subscribers have access to the code used in this project as well as the micro-detailed configuration information for the solution. To learn about subscription options, please contact us.