Embrace, Extend, Evolve - Making PostgreSQL Work for a Cloud Native World
summary
YugabyteDB reuses PostgreSQL’s parser/planner but replaces storage with an LSM engine and a 64‑bit hybrid logical clock avoids XID wraparound, eliminating VACUUM pain and optimizing SSDs (LSM storage eliminates VACUUM). It implements raft consensus for symmetric multi‑zone HA, supports both synchronous and asynchronous geo replication, and pushes connection pooling into the DB with a Connection Manager (in development). Migration uses open‑source Voyager (Apache 2.0). Production examples claim 2M TPS on a 300+ TB cluster approaching multi‑petabyte scale.
AI-written summary · about
chapters (11)
- 0:00Welcome
- 1:06PostgreSQL dominance
- 3:14YugabyteDB and PostgreSQL
- 6:46Extending PostgreSQL
- 13:48Evolving to distributed SQL
- 21:52Cloud-native scalability
- 27:12High availability
- 35:18Geographic distribution
- 43:04Simplified architecture
- 49:21Migrating with Voyager
- 51:04Closing
related talks
30mTechnical Comparison of YugabyteDB with Oracle and PostgreSQL
25mYugabyteDB Technical Overview
22mYugabyteDB: Discover What's New
26mPostgreSQL: Revealing Secrets Behind YugabyteDB's Industry-Leading Compatibility
47mAnatomy of a Distributed SQL Database - YugabyteDB
26mEfforts Toward Adoption and Uncovering Hidden Features
22mMigrating from Monolithic to Cloud Native Distributed SQL Database with Zero Downtime
31mLatest Developments / Japan Business and Product Updates