Commit Latency vs Throughput
Commit Latency vs Throughput
Analysis of how commit latency scales with achieved throughput.
Values shown as mean ± standard deviation across seeds.
num_tables = 1
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.47 ± 0.00 | 1.53 ± 0.43 | 5.92 ± 1.73 | 1.1 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.48 ± 0.00 | 2.62 ± 0.59 | 12.61 ± 2.41 | 1.2 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.49 ± 0.00 | 5.46 ± 0.67 | 22.45 ± 4.54 | 1.3 |
| 1.6 ± 0.0 | 99.9 ± 0.0 | 0.52 ± 0.00 | 10.25 ± 0.37 | 44.05 ± 3.15 | 1.7 |
| 2.0 ± 0.0 | 99.6 ± 0.1 | 0.56 ± 0.01 | 11.43 ± 1.61 | 55.64 ± 6.38 | 2.0 |
| 2.7 ± 0.0 | 97.7 ± 0.1 | 0.72 ± 0.01 | 13.16 ± 1.45 | 69.59 ± 7.18 | 2.6 |
| 3.6 ± 0.0 | 86.6 ± 0.4 | 0.96 ± 0.01 | 7.20 ± 0.94 | 68.20 ± 4.53 | 3.6 |
| 4.5 ± 0.0 | 54.7 ± 0.4 | 1.29 ± 0.01 | 2.82 ± 0.01 | 28.46 ± 4.79 | 4.6 |
| 5.1 ± 0.0 | 31.0 ± 0.1 | 1.48 ± 0.01 | 2.78 ± 0.01 | 2.96 ± 0.01 | 5.2 |
| 5.6 ± 0.0 | 13.6 ± 0.1 | 1.53 ± 0.01 | 2.76 ± 0.01 | 2.92 ± 0.01 | 5.4 |
num_tables = 2
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.47 ± 0.00 | 1.22 ± 0.27 | 4.55 ± 1.69 | 1.0 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.48 ± 0.00 | 2.77 ± 0.76 | 10.73 ± 2.64 | 1.2 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.49 ± 0.00 | 5.37 ± 0.36 | 22.30 ± 2.35 | 1.3 |
| 1.6 ± 0.0 | 100.0 ± 0.0 | 0.51 ± 0.00 | 9.38 ± 1.02 | 43.18 ± 3.36 | 1.6 |
| 2.1 ± 0.0 | 100.0 ± 0.0 | 0.53 ± 0.00 | 12.35 ± 0.85 | 54.16 ± 7.23 | 1.8 |
| 2.7 ± 0.0 | 99.9 ± 0.1 | 0.57 ± 0.00 | 16.47 ± 1.37 | 77.38 ± 10.37 | 2.1 |
| 4.0 ± 0.0 | 98.0 ± 0.2 | 0.73 ± 0.01 | 20.76 ± 2.45 | 96.45 ± 1.96 | 3.0 |
| 6.1 ± 0.0 | 73.8 ± 0.3 | 1.18 ± 0.01 | 2.47 ± 0.02 | 42.70 ± 4.21 | 4.9 |
| 6.8 ± 0.0 | 41.4 ± 0.0 | 1.42 ± 0.00 | 2.51 ± 0.01 | 2.80 ± 0.01 | 5.6 |
| 7.2 ± 0.0 | 17.7 ± 0.0 | 1.49 ± 0.02 | 2.56 ± 0.00 | 2.76 ± 0.01 | 5.8 |
num_tables = 5
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.47 ± 0.00 | 1.63 ± 0.28 | 4.81 ± 1.14 | 1.1 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.48 ± 0.00 | 2.52 ± 0.65 | 10.17 ± 1.37 | 1.2 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.49 ± 0.00 | 5.19 ± 1.32 | 21.72 ± 3.49 | 1.3 |
| 1.7 ± 0.0 | 100.0 ± 0.0 | 0.51 ± 0.00 | 9.70 ± 0.65 | 43.36 ± 5.98 | 1.5 |
| 2.1 ± 0.0 | 100.0 ± 0.0 | 0.52 ± 0.00 | 12.34 ± 1.54 | 58.21 ± 1.95 | 1.7 |
| 2.7 ± 0.0 | 100.0 ± 0.0 | 0.54 ± 0.00 | 16.28 ± 1.25 | 74.41 ± 7.47 | 1.9 |
| 4.1 ± 0.1 | 99.9 ± 0.0 | 0.59 ± 0.00 | 25.25 ± 1.59 | 117.80 ± 5.56 | 2.4 |
| 7.9 ± 0.0 | 96.0 ± 0.2 | 0.81 ± 0.01 | 28.31 ± 2.21 | 187.80 ± 11.25 | 3.7 |
| 12.0 ± 0.0 | 73.3 ± 0.1 | 1.11 ± 0.01 | 2.09 ± 0.01 | 22.24 ± 1.24 | 5.1 |
| 15.0 ± 0.0 | 36.8 ± 0.1 | 1.47 ± 0.00 | 2.32 ± 0.01 | 2.56 ± 0.01 | 6.1 |
num_tables = 10
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.45 ± 0.00 | 1.30 ± 0.83 | 4.72 ± 0.97 | 1.0 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.47 ± 0.00 | 2.54 ± 0.60 | 11.61 ± 1.16 | 1.1 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.48 ± 0.00 | 4.79 ± 1.04 | 21.70 ± 2.38 | 1.3 |
| 1.6 ± 0.0 | 100.0 ± 0.0 | 0.51 ± 0.00 | 8.70 ± 1.21 | 43.57 ± 4.26 | 1.5 |
| 2.0 ± 0.0 | 100.0 ± 0.0 | 0.52 ± 0.00 | 11.43 ± 1.35 | 54.34 ± 4.14 | 1.6 |
| 2.7 ± 0.0 | 100.0 ± 0.0 | 0.54 ± 0.00 | 16.30 ± 1.79 | 75.32 ± 3.48 | 1.8 |
| 4.1 ± 0.0 | 100.0 ± 0.0 | 0.57 ± 0.00 | 23.79 ± 1.31 | 111.21 ± 2.89 | 2.2 |
| 8.1 ± 0.0 | 98.7 ± 0.1 | 0.67 ± 0.00 | 45.21 ± 1.26 | 230.58 ± 6.72 | 3.4 |
| 13.6 ± 0.0 | 82.9 ± 0.1 | 0.89 ± 0.00 | 1.85 ± 0.00 | 106.09 ± 7.65 | 4.6 |
| 17.8 ± 0.0 | 44.0 ± 0.1 | 1.12 ± 0.00 | 1.83 ± 0.00 | 2.05 ± 0.00 | 5.8 |
num_tables = 20
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.44 ± 0.01 | 0.93 ± 0.32 | 4.83 ± 1.17 | 1.0 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.46 ± 0.00 | 2.55 ± 0.50 | 11.22 ± 2.78 | 1.1 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.48 ± 0.00 | 4.85 ± 0.83 | 23.87 ± 4.80 | 1.2 |
| 1.6 ± 0.0 | 100.0 ± 0.0 | 0.50 ± 0.00 | 8.66 ± 1.35 | 42.95 ± 2.46 | 1.5 |
| 2.1 ± 0.0 | 100.0 ± 0.0 | 0.52 ± 0.00 | 11.39 ± 1.27 | 55.10 ± 7.15 | 1.6 |
| 2.7 ± 0.0 | 100.0 ± 0.0 | 0.54 ± 0.00 | 15.55 ± 2.68 | 77.79 ± 6.71 | 1.8 |
| 4.1 ± 0.0 | 100.0 ± 0.0 | 0.57 ± 0.00 | 25.76 ± 1.52 | 121.12 ± 4.99 | 2.1 |
| 8.1 ± 0.0 | 99.2 ± 0.1 | 0.64 ± 0.00 | 46.06 ± 2.35 | 233.22 ± 15.46 | 3.1 |
| 13.9 ± 0.0 | 85.5 ± 0.1 | 0.81 ± 0.00 | 26.68 ± 0.65 | 210.16 ± 10.79 | 4.5 |
| 18.3 ± 0.1 | 45.2 ± 0.1 | 1.01 ± 0.00 | 1.60 ± 0.01 | 1.86 ± 0.01 | 5.7 |
num_tables = 50
| Throughput (c/s) | Success Rate (%) | P50 Latency (s) | P95 Latency (s) | P99 Latency (s) | Mean Retries |
|---|
| 0.2 ± 0.0 | 100.0 ± 0.0 | 0.37 ± 0.02 | 0.64 ± 0.15 | 3.73 ± 1.25 | 1.0 |
| 0.4 ± 0.0 | 100.0 ± 0.0 | 0.45 ± 0.00 | 2.18 ± 0.99 | 11.36 ± 2.63 | 1.1 |
| 0.8 ± 0.0 | 100.0 ± 0.0 | 0.46 ± 0.00 | 4.90 ± 0.56 | 20.06 ± 3.20 | 1.2 |
| 1.6 ± 0.0 | 100.0 ± 0.0 | 0.49 ± 0.00 | 9.55 ± 0.77 | 45.95 ± 3.82 | 1.4 |
| 2.0 ± 0.0 | 100.0 ± 0.0 | 0.51 ± 0.00 | 11.69 ± 1.18 | 55.72 ± 4.82 | 1.5 |
| 2.7 ± 0.0 | 100.0 ± 0.0 | 0.53 ± 0.00 | 14.54 ± 1.64 | 76.20 ± 7.34 | 1.7 |
| 4.1 ± 0.0 | 100.0 ± 0.0 | 0.56 ± 0.00 | 23.09 ± 1.60 | 113.34 ± 5.61 | 2.0 |
| 8.1 ± 0.0 | 99.4 ± 0.0 | 0.63 ± 0.00 | 49.75 ± 4.44 | 236.72 ± 14.74 | 3.0 |
| 14.1 ± 0.0 | 86.5 ± 0.3 | 0.76 ± 0.00 | 44.13 ± 2.93 | 316.54 ± 9.22 | 4.4 |
| 18.5 ± 0.1 | 45.7 ± 0.2 | 0.97 ± 0.00 | 1.50 ± 0.00 | 72.69 ± 5.70 | 5.8 |
Notes
- Latencies reported in seconds (converted from milliseconds)
- Values shown as mean ± standard deviation across multiple seeds
- Throughput = commits per second during steady-state window
- Success rate = percentage of transactions that committed successfully
- Mean retries = average number of retry attempts per committed transaction
