Performance Benchmarks
Performance characteristics of the Native Content Relationships engine at scale. Benchmarks use deterministic test datasets and SQL-native validation.
Test environment
- Dataset size: 100,000 – 1,000,000 relationship rows
- Storage engine: InnoDB
- Indexing: Composite covering index (
type_lookup) - WordPress: 6.x · PHP: 7.4+
- Object cache: Disabled (baseline)
- MySQL buffer pool: Warmed before final run
Latency metrics
| Operation | 100k rows | 1.0M rows |
|---|---|---|
| Point lookup (mean) | 0.49 ms | 1.00 ms |
| Point lookup (P95) | 0.85 ms | 2.73 ms |
| Covering index mean | 0.22 ms | 0.61 ms |
| Covering index P95 | 1.25 ms | 3.42 ms |
| Full graph scan | ~7.2 s | ~64.2 s |
Variation depends on buffer pool state and cache warm-up.
Resource efficiency
Memory
- Peak memory delta (1.0M rows): ~2.21 MB
- Max observed: < 5 MB
- Scaling: Independent of dataset size
Suitable for shared hosting and restricted environments.
Database strategy
Covering index: KEY type_lookup (type, from_id, to_id) — index-only lookups for common queries; query time growth O(log n). Under typical workloads, query latency stays sub-2ms at 1M rows.
Methodology
Benchmarks run via the benchmarks/performance-report.php utility: deterministic graph generation, buffer pool warming, mean latency over iterations, peak memory via memory_get_peak_usage().
TIP
For high-availability setups, monitor P95 query latency during full integrity scans.