Meta-Data Caching and Record Store Scalability
Talkat FoundationDB Summit 2019 · on Monday, 2019-11-18· at 13:12· Grand Ballroom 5, Marriott Marquis San Diego Marina·
summary
A Record Layer cache uses a tiny, proxy-served metadata version key (12 bytes) added in FoundationDB 6.1 and set only via a version stamped value to detect transactional schema changes without hitting the same storage key on every request. The implementation, the metadata version stamp store state cache, invalidates and refreshes cached schema atomically with commits, cutting read hotspot load and letting a single record store scale past the previous single-key limit; multi-tenant conflict-range tweaks handle remaining edge cases.
AI-written summary · about
related talks
05mLightning Talk: Entity Store: A FoundationDB Layer for Versioned...
22mHow FoundationDB powers SnowflakeDB's metadata!
26mFuture of FoundationDB Storage Engines
37mTechnical Overview of FoundationDB
21mCore Data Custom Store
59mSingleStore: Storage Metadata for Modern Cloud Databases
01h 01mClosing Keynote
46mFoundationDB: The Best-Kept Secret of New Distributed Architectures!
