AliSQL Enhances MySQL with Vector and DuckDB Engines

AliSQL Enhances MySQL with Vector and DuckDB Engines

AliSQL, a branch of MySQL developed by Alibaba, integrates advanced features such as vector processing and DuckDB engines. It is designed to optimise performance and stability for large-scale applications, making it a robust choice for enterprise environments.

AliSQL is a specialised version of MySQL, forked and maintained by Alibaba Group, tailored to meet the demands of large-scale applications. It incorporates various performance optimisations and stability improvements, making it a reliable choice for enterprise-level deployments. One of the standout features of AliSQL is its integration with DuckDB as a native storage engine, allowing users to operate DuckDB with the same ease as MySQL. This integration facilitates the rapid deployment of DuckDB service nodes, enabling lightweight analytical capabilities.

In addition to DuckDB integration, AliSQL supports enterprise-grade vector processing, accommodating up to 16,383 dimensions. This feature is particularly beneficial for AI-driven applications, such as semantic search and recommendation systems, as it utilises a highly optimised HNSW algorithm for high-performance Approximate Nearest Neighbor (ANN) search. These capabilities are accessible through standard SQL interfaces, making it easier for developers to build sophisticated applications without needing to learn new query languages.

Technical Details

AliSQL's integration with DuckDB as a native storage engine is a significant enhancement, providing users with the ability to leverage DuckDB's analytical capabilities seamlessly. This integration allows for the rapid deployment of DuckDB nodes, which is particularly useful for applications requiring lightweight analytics. The vector processing support in AliSQL is another key feature, enabling high-dimensional data processing with up to 16,383 dimensions. This is achieved through the use of an optimised HNSW algorithm, which enhances the performance of ANN searches, crucial for applications like semantic search and recommendation systems.

Planned optimisations in AliSQL include improvements in DDL operations, crash recovery, and replication. The DDL optimisation aims to provide a faster, safer, and lighter experience through innovations such as enhanced Instant DDL and parallel B+tree construction. These improvements are expected to significantly enhance schema change efficiency and reduce replication lag. Additionally, AliSQL plans to optimise the crash recovery path to accelerate instance startup and restore services quickly, while replication throughput is set to be boosted through various specialised optimisations.

Getting Started

AliSQL became an open-source project in December 2025 and is actively maintained by engineers at Alibaba Group. It is licensed under the GPL-2.0 license, similar to MySQL, and contributions from the community are encouraged. For those interested in using AliSQL, the project provides build instructions and options, including specific support for DuckDB. Users can report bugs and request features through the GitHub Issues page, ensuring that the project continues to evolve with community input.

Story based on discussion on Hacker News.