This test was performed from a single client. The tests were run with different warehouses like 1000, 1500, 2000, 2500, and 2700 warehouses by running the following command and the results were captured.
# ./tpccbenchmark --execute=true --nodes=<List_of_IP_address_of_database_nodes> --warehouses=<number_of_warehouses> --warmup-time-secs 300
| TPCC Metrics | PowerFlex Metrics | |||||
Warehouse | TPM-C | Efficiency | Throughput (requests/sec) | Read Latency (ms) | Read IOPS | Write Latency (ms). | Write IOPS |
1,000 | 12,857 | 99.98 % | 476 | 0.31 | 2,882 | 0.37 | 1,419 |
1,500 | 19,258 | 99.83 % | 713 | 0.30 | 5,790 | 0.42 | 1,484 |
2,000 | 25,683 | 99.85 % | 950 | 0.30 | 9,230 | 0.47 | 1,571 |
2,500 | 32,082 | 99.79 % | 1,188 | 0.29 | 13,020 | 0.48 | 1,693 |
2,700 | 34,652 | 99.80 % | 1,282 | 0.29 | 14,460 | 0.51 | 1,788 |
It was observed that the average CPU utilization on each database node was below 80%. As the warehouses were scaled, TPM-C, throughput, PowerFlex read and write IOPS values increased linearly.