MySQL Database Service, HeatWave, and Autopilot
summary
Massively parallel in‑memory query accelerator HeatWave integrates with MySQL Database Service and uses machine‑learning MySQL Autopilot to auto‑size clusters (sampling <0.1% and 95% accuracy), choose in‑memory partition keys, and improve query plans. Claims include 6.3s vs 2.5 hours on 4TB TPC‑H (about 1400x faster than Aurora) and 44× price‑performance vs Snowflake at 30TB; hybrid on‑premise replication works and published scripts reproduce the benchmarks.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 0:25MySQL Database Service
- 2:36HeatWave
- 3:25Integration
- 5:58Massive Parallel Architecture
- 7:53Data Management
- 9:20Hybrid Deployment
- 10:32Cloud Automation
- 11:37Machine Learning Based Automation
- 12:27Machine Learning Models
- 14:00Autopilot Features
- 16:37HeatWave Partition
- 18:20Kodi Execution
- 19:59Summary
- 20:27Benchmarks
- 25:13Mixed workloads
- 26:16Red 3i
- 26:48Tetris
- 27:28Tamara
- 27:54Conclusion
- 29:21Handson Lab
related talks
19mDo You Believe in Magic?
44mOracle Cloud Your Way: Multicloud, Dedicated, Alloy, Sovereign
06mScaling your applications with Cloud SQL for MySQL
46mPower the next generation of AI apps with databases at scale anywhere
55mThe Future of MySQL
53mMySQL DevOps
32mData @Scale Seattle- Eric Hwang introduced by Ganapathy Krishnamoorthy
41mHTAP with MariaDB Platform