Vector Indexes in MariaDB, A Performance Comparison

Abstract

As AI-driven applications and Retrieval-Augmented Generation (RAG) systems continue to grow in complexity, efficient data retrieval is becoming a critical factor in performance and scalability. Vector datatypes, key to similarity search and embeddings, are now being adopted across various databases, including MariaDB.
 This presentation will introduce the audience to vector datatypes in MariaDB, and demonstrate their capabilities for storing and querying high-dimensional data. We will then compare MariaDB’s vector performance against other technologies such as PostgreSQL (pgvector), redisearch, and specialized vector databases like Qdrant and Weaviate.

Speaker: Alex Hanshaw – VP Engineering @MariaDB