BemiDB is a Postgres read replica optimized for analytics, featuring a single binary that replicates data in a compressed columnar format. It boasts up to 2000x faster query performance than Postgres, supports seamless integration with Postgres-compatible tools, and stores data either locally or on S3. BemiDB utilizes the DuckDB query engine and the Iceberg table format for efficient data handling and is available as an open-source project under the AGPL-3.0 License.

5m read timeFrom github.com
Post cover image
Table of contents
ContentsHighlightsUse casesQuickstartConfigurationArchitectureFuture roadmapBenchmarkDevelopmentAlternativesLicense

Sort: