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 promisingOne-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
The optimizer delivers more efficient execution plans due to its support for cost-based vectorized path search.
Vectorized execution engines process tuples in batches, achieving performance improvements of hundreds or even thousands of times.
Using TOAST technology to implement a columnar storage engine, achieving a data compression ratio improvement of 3-5 times.
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.
GammaDB is
Comming Soon!
Currently, GammaDB does not yet possess product-level capabilities, and more robust functions are being progressively enhanced.