Alibaba's MySQL Branch with DuckDB Analytics and Native Vector Search
Based on MySQL 8.0.44 and maintained by Alibaba Cloud Database Team
Features • Quick Start • Docs • Roadmap • Contributing
AliSQL is an open-source MySQL branch maintained by Alibaba Cloud Database Team. This release is based on MySQL 8.0.44 and adds a DuckDB analytical storage engine, HNSW vector indexes, Native Flashback, binlog realtime replication (BRR), and binlog optimizations while retaining the MySQL client protocol.
The 2026-10-16 update introduces BRR and upgrades the embedded DuckDB snapshot to v1.5.2-dev736, with optional AUTO_INCREMENT and background checkpoints.
|
The included TPC-H SF100 results show more than 200x speedup over InnoDB for several queries in that test environment |
The |
Applications connect through the MySQL protocol and continue to use MySQL clients and drivers |
| Feature | Description | Status |
|---|---|---|
| DuckDB Storage Engine | Columnar execution and storage for analytical tables, with automatic compression | Available |
| Vector Index (VIDX) | Native vector storage & ANN search with HNSW, supports COSINE & EUCLIDEAN distance | Available |
| Native Flashback | Query historical InnoDB data with AS OF TIMESTAMP and retained undo snapshots |
Available |
| Large TX Optimization | Binlog Cache Free Flush reduces commit-time I/O amplification for large InnoDB transactions | Available* |
| Binlog Durability | Persist Binlog Into Redo V2 reduces synchronous binlog I/O while retaining crash recovery | Available |
| Binlog Realtime Replication (BRR) | Transmit eligible large transactions before commit and optionally apply supported DDL and transactions early on replicas | Available, opt-in |
| DDL Optimization | Instant DDL, parallel B+tree construction, non-blocking locks | Planned |
| RTO Optimization | Accelerated crash recovery for faster instance startup | Planned |
* Free Flush targets eligible InnoDB transactions while DuckDB mode is inactive. With DuckDB mode enabled, commit uses the dedicated DuckDB path. BRR reuses Free Flush cache files; this does not by itself imply a Free Flush commit.
# Clone the repository git clone https://github.com/alibaba/AliSQL.git cd AliSQL # Build (release mode) sh build.sh -t release -d ~/alisql # Install make install
Step-by-step guide: How to set up a DuckDB node
# Initialize data directory ~/alisql/bin/mysqld --initialize-insecure --datadir=~/alisql/data # Start the server with DuckDB enabled for the example below ~/alisql/bin/mysqld --datadir=~/alisql/data --duckdb_mode=ON
-- Create an analytical table with DuckDB engine CREATE TABLE sales_analytics ( sale_date DATE, product_id INT, revenue DECIMAL(10,2), quantity INT ) ENGINE=DuckDB; -- Run an analytical aggregation through DuckDB SELECT DATE_FORMAT(sale_date, '%Y-%m') as month, SUM(revenue) as total_revenue, COUNT(*) as transactions FROM sales_analytics GROUP BY month ORDER BY total_revenue DESC;
-- Vector features are disabled by default and vector indexes require RC. SET GLOBAL vidx_disabled = OFF; SET SESSION transaction_isolation = 'READ-COMMITTED'; -- Create a table with a vector column CREATE TABLE embeddings ( id INT PRIMARY KEY, content TEXT, embedding VECTOR(3) ) ENGINE=InnoDB; INSERT INTO embeddings VALUES (1, 'first document', VEC_FROMTEXT('[0.1,0.2,0.3]')), (2, 'second document', VEC_FROMTEXT('[0.2,0.1,0.4]')); -- Create HNSW index for fast ANN search CREATE VECTOR INDEX idx_embedding ON embeddings(embedding) DISTANCE=COSINE; -- Find similar items using cosine distance SELECT id, content, VEC_DISTANCE_COSINE( embedding, VEC_FROMTEXT('[0.1,0.2,0.3]') ) AS distance FROM embeddings ORDER BY distance LIMIT 10;
| Option | Description | Default |
|---|---|---|
-t release|debug |
Build type | debug |
-d <dir> |
Installation directory | /usr/local/alisql when writable; otherwise $HOME/alisql |
-g asan|tsan |
Enable sanitizer (memory/thread) | disabled |
-c |
Enable code coverage (gcov) | disabled |
Prerequisites: CMake 3.x+, Python 3, GCC 7+ or Clang 5+
Q4 2025 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[x] DuckDB Storage Engine [x] Vector Index (VIDX) [x] 8.0.44 Release
2026 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[x] Native Flashback [x] Transaction & Binlog [ ] DDL / RTO
- AS OF TIMESTAMP - Binlog in Redo V2 - Instant DDL
- Undo snapshots - Free Flush - Fast Crash Recovery
- Realtime Replication
| Document | Description |
|---|---|
| DuckDB Integration Guide | Architecture, compatibility, replication, and performance reference |
| Vector Index Guide | Native vector storage and ANN search |
| Native Flashback Guide | Historical InnoDB queries and recovery |
| Binlog in Redo Guide | Redo-backed binlog persistence and fallback rules |
| Binlog Cache Free Flush Guide | Large-transaction commit path, configuration, and restrictions |
| Binlog Realtime Replication Guide | BRR transmission, early apply, configuration, monitoring, and limitations |
| 2026-10-16 Release Notes | BRR, DuckDB updates, and integration fixes |
| Release Notes | What's new in AliSQL 8.0.44 |
| Setup DuckDB Node | Quick setup guide for analytics |
External Resources:
Alibaba Cloud RDS MySQL provides managed versions of several AliSQL features. RDS manages engine releases, topology, data synchronization, backup, monitoring, and support. Supported versions and parameter defaults can differ from this repository; use the product documentation when operating an RDS instance.
| Feature | RDS MySQL product documentation |
|---|---|
| DuckDB analytical instances | English / 中文 |
| Vector storage | English / 中文 |
| Native Flashback | English / 中文 |
| Binlog Cache Free Flush | English / 中文 |
| Binlog in Redo | English / 中文 |
The local feature guides note behavior that differs between this source tree and RDS MySQL.
AliSQL has been open source since August 2016 and is actively maintained by Alibaba Cloud Database Team. The current 8.0.44 feature release continues that open-source line.
Bug reports, documentation fixes, and code changes are welcome.
- Fork the repository.
- Create a feature branch (
git checkout -b feature/my-change). - Commit the change (
git commit -m 'Describe the change'). - Push the branch (
git push origin feature/my-change). - Open a pull request.
For bugs and feature requests, please use GitHub Issues.
RDSAI CLI is a command-line client for AliSQL and MySQL. It accepts natural-language requests for SQL generation, execution-plan analysis, and query diagnosis.
# Install curl -LsSf https://raw.githubusercontent.com/aliyun/rdsai-cli/main/install.sh | sh # Connect and ask in natural language rdsai --host localhost -u root -p secret -D mydb mysql> analyze index usage on users table mysql> show me slow queries from the last hour mysql> why this query is slow: SELECT * FROM users WHERE name LIKE '%john%'
Key features:
- Natural language to SQL conversion (English/中文)
- Query optimization and diagnostics
- Execution plan analysis with
Ctrl+E - Multi-model LLM support (Qwen, OpenAI, DeepSeek, Anthropic, etc.)
- Performance benchmarking with automated analysis
Project and installation instructions
|
GitHub Issues For bug reports & feature requests |
Alibaba Cloud RDS Managed DuckDB analytical instances |
AliSQL is licensed under GPL-2.0, the same license as MySQL.
See the LICENSE file for details.
Maintained by Alibaba Cloud Database Team
