Technical Overview of FoundationDB
summary
FoundationDB provides strict serializability across a global keyspace by assigning committed transactions monotonic versions from a single master, serving reads with a single hop to storage servers and committing writes through about four hops (~4–5 ms in ideal conditions). The design composes many small key‑value storage servers into one cluster achieving roughly 90% of theoretical scalability, enforces a five-second transaction limit, and relies on a deterministic single‑process simulator with randomized failure tests to validate correctness.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:47FoundationDB strengths
- 4:04Distributed databases 101
- 8:03Stateful components
- 12:02Consistency
- 14:28API
- 14:57Reads
- 18:43Commits
- 25:01Failures
related talks
46mFoundationDB: The Best-Kept Secret of New Distributed Architectures!
50mBuilding the Next-Generation Backend on FoundationDB
34mFoundationDB, the Black Knight
25mBuild Anything with FoundationDB
16mFoundationDB and Node.js: Year in Production
21mEvaluation of FoundationDB in Financi...
24mHow We Saved 5x Migrating from Aurora to FoundationDB
15mTime Series and FoundationDB: Millions of Writes/s and 10x Compression in 2000 Lines of Go