Benchmarks / results
Performance with the conditions attached.
These are user-reported local development measurements, presented as a specialized raw-ingestion comparison—not a universal claim about production performance.
Observed comparison · docs/sec
No animation required to understand resultPeak rate and full-run average are separate metrics. The average is calculated from 10 million documents over 36 seconds.
Peak write rate
12×
Peak-rate ratio: 300K observed LioranDB writes/sec versus 25K reported MongoDB writes/sec.
End-to-end ingestion
11.1×
36 seconds versus 400 seconds for this 10M-document local workload.
Reported image size
72.6% smaller
Approximate 45 MB versus 164 MB reported image sizes; separate from the Windows benchmark environment.
Configuration
The workload is intentionally narrow.
Disabling secondary and text indexes reduces write amplification. Simple username/age documents do not represent arbitrary application data.
| Setting | LioranDB | MongoDB comparison |
|---|---|---|
| Documents | 10,000,000 | 10,000,000 |
| Document shape | Simple username and age | Not further recorded |
| Batch size | 512 | 512 reported |
| Indexes | Secondary and text disabled | Index semantics not recorded |
| Node / partitions | 1 internal node / 1 partition | Not recorded |
| Durability | WAL and checkpointing enabled | Write concern and fsync behavior not recorded |
| Memory | ≈3.5 GiB observed | Not measured in supplied comparison |