litendb 0.0.14

Liten Database

Concept

The Liten database constructs a tensor-based data structure that uses indices to link dimensional data with corresponding metrics. This architecture transforms traditional tabular relationships into highly efficient index lookups, resulting in dramatically faster query execution.

Implementation Detail

A GitHub module provides a method that automatically creates data tensors during data persistence. By generating these tensors as data is stored, the system ensures query acceleration at the source, minimizing runtime computation overhead.

Performance Benefits

Many query operations become direct index lookups rather than complex joins. Multi-dimensional tensors simplify query logic and boost performance. The approach enables high-speed analytical queries across large datasets.

Acceleration Technology

• Novel distributed tensor-based engine with multi-dimensional hierarchical data models
• Faster execution compared to traditional relational models
• No joins, fast aggregates, and simpler query plans
• Further acceleration with GPU and FPGA hardware tensors
• Open lakehouse storage, enabling interoperability with other analytics tools