2026-04-24 · 読了目安 1 分
UUID v4 vs. Sequential IDs
Choose randomness for distribution and privacy, or sequence for storage locality.
UUID v4 is randomly generated, making it hard to guess and great for distributed systems.
Sequential IDs are easy to index and sort, but reveal your database volume and are predictable.
この比較の見方
Selection depends on whether you need to hide volume (security) or optimize index inserts (performance).
| アプローチ | データ処理 | 一般的な速度 | 適した用途 |
|---|---|---|---|
| UUID v4 | 128-bit random; no coordination needed between servers | Slower for B-tree index inserts due to fragmentation | Distributed systems, public-facing IDs, privacy-sensitive records |
| Sequential (Auto-increment) | Typically 32/64-bit integer; requires central counter | Fastest for database inserts; compact storage | Internal tables, low-concurrency systems, storage-optimized apps |
要点
- Use UUID v4 for public IDs to prevent ID enumeration attacks.
- Use sequential IDs for high-volume internal logging where insert speed is the bottleneck.