One-Click Transform PostgreSQL to Native Analytical Database

GammaDB is a PostgreSQL extension with self-developed vectorized execution and column storage, enhancing performance without needing third-party analytical engines.

The architecture of GammaDB

Performance Boost 1000x!

Native Analytical PostgreSQL

The future of AI and distributed capabilities looks promising

One-click transformation

Leveraging PostgreSQL's extension capabilities, achieve one-click transformation into an analytical engine, without relying on any third-party databases.

Decoupling Storage and Compute

Leveraging S3 to offload data to object storage, achieving multi-dimensional separation of compute, transactions, metadata, and data.

AI and distributed execution

The capabilities of AI in DB and DB in AI are already on our roadmap, and the ability for distributed execution is also under development. The architecture of GammaDB will continue to be refined and improved.

Components

Fully self-developed Optimization, Execution, and Storage Engines

Comprehensively enhancing PostgreSQL's capabilities

Cost-based Vectorized Optimization

The optimizer delivers more efficient execution plans due to its support for cost-based vectorized path search.

Vectorized Execution

Vectorized execution engines process tuples in batches, achieving performance improvements of hundreds or even thousands of times.

Columar Storage Engine

Using TOAST technology to implement a columnar storage engine, achieving a data compression ratio improvement of 3-5 times.

Columnar Index For HOT

By building a columnar index as a covering index on a heap-organized table, AP-style SQL queries experience a 1000x performance improvement while TP performance remains essentially unchanged.

??.??K
Downloads
??.??K
Stars
??.??K
Forks
??.??K
Users

GammaDB is  
Comming Soon!

Currently, GammaDB does not yet possess product-level capabilities, and more robust functions are being progressively enhanced.