Introducing Kunlun Distributed Database Cluster
summary
Kunlun claims a working distributed DBMS that wires PostgreSQL computing nodes to MySQL storage shards, implements two-phase commit for multi-shard transactions, uses MySQL group replication for shard HA, and adds a global deadlock detector and reliable DDL sync so all compute nodes keep local metadata. The project fixes MySQL 8.0 XA bugs, reports up to ~50% higher throughput and ~50% lower latency in a PoC, and v0.7 is listed as poc ready.
AI-written summary · about
chapters (13)
- 0:00Start
- 2:33Hardware infrastructure for computing
- 4:56Future DBMS Requirement
- 5:42Existing solutions --- common
- 10:36Pain points
- 11:56Best approach --- Distributed DBMS
- 13:38Kunlun Architecture
- 15:22Kunlun Basic
- 22:38DML & transactions in
- 23:41Kunlun Indepth ---Computing nodes
- 36:09Kunlun Indepth -- Storage shards
- 38:11Kunlun Development Progress
- 39:16Kunlun Development Plan
related talks
31m#PostgreSQL to #MySQL: A DBA Perspective
42mPostgreSQL Distributed & Secure Database Ecosystem Building
46mDatabase Sharding the Right Way Easy, Reliable, and Open source
24mComparing the High Availability of MySQL with Modern Distributed SQL Databases
49mOnline Database Reliability, Performance, and Consistency Engineering
30mThe database trends that are transforming your database infrastructure forever
31mOpen Source Databases on Kubernetes: Best Practices
55mThe Future of MySQL