Why choose Apache Cloudberry™

Apache Cloudberry™ is a Greenplum® fork managed by Apache Software Foundation. Apache Cloudberry™ version 2.x is built on the modern PostgreSQL 14.x, while Greenplum® is based on PostgreSQL 9.4 whose support was discontinued in February, 2020. The upcoming major Apache Cloudberry™ 3.x release will upgrade to PostgreSQL 16.

Migrating from Greenplum® to Apache Cloudberry™ is straightforward: Apache Cloudberry™ is almost fully compatible with Greenplum®, so you do not need to retrain your team or rewrite your applications. For the full list of incompatible changes, which are mostly minor and isolated, see this article. These changes affect very few workflows.

Benefits Apache Cloudberry™Benefits Apache Cloudberry™

Faster analyticsFaster analytics

To a large extent, the speed of analytical database engines depends on their data storage format. Apache Cloudberry™ uses a hybrid data storage format called PAX (Partition Attributes Across). It combines the convenience of row-oriented tables with the fast read speeds of column-oriented tables, automatically skips irrelevant sections, and supports modern compression codecs.

Query optimizationQuery optimization

Apache Cloudberry™ features native query optimization tools:

• RuntimeFilter: Filters out excessive tuples before JOIN operations using bloom filters.
• Aggregation Pushdown: Performs aggregations closer to the data, reducing the amount of network-transmitted data.
• Parallel query execution while dynamically adjusting the number of workers.
• New BRIN index type that supports the multi-minmax and bloom options.
• PostgreSQL 14 query optimizer, upgraded and improved.

Near-real-time analytics and AI integrationNear-real-time analytics and AI integration

• Dynamic tables: Query results are updated on a schedule.
• Incremental materialized views: Only modified data is updated.
• Automatic matrialized views (AQUMV): Optimizer automatically substitutes tables with the relevant materialized views when generating a query plan.
• Directory tables: Manage unstructured data residing in a local or object storage directly via SQL.
• MCP server: Manage your database using AI agents and LLM tools.

Corporate securityCorporate security

• Password policies based on profiles, including lockouts after multiple failed login attempts.
• Row level security to restrict access to specific rows without view-based workarounds.
• SCRAM-SHA-256, channel binding, TCP/IP encryption with GSSAPI, and trusted extensions.

Vendor-neutral governance modelVendor-neutral governance model

• The project was developed in the Apache Software Foundation Incubator which guarantees vendor neutrality.
• Development decisions are taken by the Project Management Committee (PMC), which includes developers from various companies and countries.
• Active community and fast development.