Efforts Toward Adoption and Uncovering Hidden Features
summary
NTT Data pilots YugabyteDB to build a spatio-temporal pipeline: Spark pre-processes mobility data and inserts it into YugabyteDB using PostGIS support. Tests show a packed-row storage format yields much higher insert+read throughput, and the cloud migration aims for ~100k read QPS, ~100k writes, tens of TB of data, 1s tail latency and zero planned downtime; noted Yugabyte engine features include a rule-based optimizer, Postgres-compatible read-committed semantics, work-in-progress atomic DDL, built-in connection pooling, and bitmap scans.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:25NTT Data’s distributed DB efforts
- 1:32YB Database for spatiotemporal data
- 2:59Cloud drift case study
- 5:23YB Database’s 3 motivations
- 6:28YB Database’s hidden features
- 6:49Rule-based optimizer traits
- 9:23Read committed benefits
- 11:48Migration benefits of isolation
- 12:50AtomicDDL for ops
- 13:32Packed row storage
- 16:47Packed row flexibility
- 17:10PostgreSQL extensions in YB DB
- 19:22Custom extensions for YB DB
- 20:263 standout YB Database features
- 24:20Summary
related talks
51mEmbrace, Extend, Evolve - Making PostgreSQL Work for a Cloud Native World
22mYugabyteDB: Discover What's New
25mYugabyteDB Technical Overview
33mThe Future of YugabyteDB : Vision, Innovation & Roadmap
27mWelcome to DSS Asia 2022 Japanese Track
26mPostgreSQL: Revealing Secrets Behind YugabyteDB's Industry-Leading Compatibility
27mSparking our way to YugabyteDB: Our journey from ScyllaDB to
22mMigrating from Monolithic to Cloud Native Distributed SQL Database with Zero Downtime