The database-on-S3 debate: Caches and slow storage are nothing new
mgill25 · x · 2026-08-29
Following the debates on WAL-on-S3 and database-on-S3, the author argues that everyone is missing the point.
First, the anxiety over whether a cache layer disqualifies it as a "database-on-S3" is misplaced. Databases for the last 40 years have always had cache layers above extremely slow storage. Whether it's old Netapp devices with RAID-5 or high-end Symmetrix arrays, databases were designed for slow storage, and caches are essential. Second, getting this to work reliably on S3 is not trivial. While one might "vibe-code" it, doing it right requires significant experience in building distributed systems.
More from Infra
- Four practical ways to optimize end-to-end AI latency — bibryam · 2026-08-29
- Analysis of 31k LLM benchmarks: within-day variation 2.8 pts — ionutvi · 2026-08-29
- Performance Optimization: Latency Reduced from 8ms to 0.87ms — DanielLockyer · 2026-08-29
- DGX Spark benchmarks: DeepSeek V4 Flash passes 900K-token prompt locally — jtsaint333 · 2026-08-29
- AutoFAB uses robot arms to take over 3D printers for 24/7 autonomous production — TinfoilTricorn · 2026-08-29
- Theory: OpenAI model was trained on victims' infrastructure schematics — Kremho · 2026-08-29