The MariaDB Foundation has announced a preview version of MariaDB 11.6 Vector. The new functionality is described as being a result of collaborative work by employees of MariaDB plc, MariaDB Foundation and contributors, particularly from Amazon AWS.

The new release is just one of a number of product updates announced by The MariaDB Foundation, including various latest stable releases in their respective long-term series, short-term series, MariaDB 11.5.2, the stable rolling release, and MariaDB 11.6.1, the Release Candidate rolling release.

The MariaDB Foundation is a separate entity from MariaDB plc, which last October announced reduction in the company's workforce by 28% and the dropping of its SkySQL and Xpand products. Kaj Arnö, CEO of the MariaDB Foundation, said recently:

"Our mission is around Adoption, Openness, and Continuity, whereas they are a for-profit startup serving paying customers."

The main highlights of the new vector search support is that the implementation is based on the industry standard HNSW (Hierarchical Navigable Small World) algorithm, and comes with syntax for easily storing vectors. In practical terms this is achieved by dedicated functions to convert a binary vector into a JSON array of numbers, and to take an array of numbers and convert it to a binary vector.

Unlike extensions for databases such as Postgre's PG Vector, the vector search functions in MariaDB are not an extension, but a core part of the native syntax of all MariaDB Server binaries. MariaDB developers have said that Much like PG Vector, MariaDB Vector will keep the LLM Large Language Model) itself separate from the binaries, and be agnostic towards the user's choice of LLM, meaning users will be able to use ChatGPT, Claude, LLaMa2 and many other LLMs with MariaDB.

The team says that all of MariaDB Vector is on the way of becoming part of standard MariaDB Server.

The creation of the vector database and vector search for MariaDB means it can compete in the market for managing multi-dimensional data as part of advanced AI applications. Until now, MySQL and MariaDB have used traditional database methods with special data types for vector storage.

More Information

MariaDB Site

Related Articles

MariaDB Cuts Workforce And Drops Products  

Azure Drops MariaDB Support, Adds Free SQL DB

MariaDB Offers Observability To Community Version

MariaDB Moves Into GeoSpatial Space

MariaDB SkySQL Improves AWS Support

MariaDB 10.4 Reaches Release Candidate Status

MariaDB Announces New Enterprise Server

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info