DBX
High-performance embedded database built on a 5-Tier Hybrid Storage architecture. Designed for HTAP (Hybrid Transactional/Analytical Processing) workloads, implemented in pure Rust.
Key Features
🏗️ Architecture
- 5-Tier Hybrid Storage — Delta → Cache → WOS → Index → ROS
- HTAP Support — Concurrent OLTP and OLAP workloads
- MVCC Transactions — Snapshot Isolation with Garbage Collection
- Columnar Cache — Apache Arrow-based analytical query optimization
🚀 Performance
- 29x faster file GET than SQLite
- GPU Acceleration — CUDA-based aggregation, filtering, joins
- SIMD Vectorization — Optimized numerical operations
- Parallel Query — Rayon-based parallel JOIN, Sort, Columnar Build
🔐 Security
- AES-256-GCM-SIV — Industry-standard encryption
- ChaCha20-Poly1305 — High-speed mobile encryption
- Key Rotation — Zero-downtime key updates
🌐 Multi-Language
- Rust — Native API
- Python — PyO3-based bindings
- C#/.NET — P/Invoke FFI
- C/C++ — Standard C API
- Node.js — Native N-API bindings
Quick Example
use dbx_core::Database;
let db = Database::open_in_memory()?;
// CRUD
db.insert("users", b"user:1", b"Alice")?;
let val = db.get("users", b"user:1")?;
// SQL
let results = db.execute_sql("SELECT * FROM users WHERE age > 25")?;
// Transactions
let tx = db.begin_transaction()?;
tx.insert("users", b"user:2", b"Bob")?;
tx.commit()?;
Table of contents
- Getting Started
- Architecture
- Benchmarks
- Language Bindings Performance
- Packages
- API Reference
- Changelog
- Examples
- Quick Start
- SQL Quick Start
- Encryption
- Compression
- Indexing
- WAL Recovery
- CRUD Operations
- SQL Reference
- Transactions
- Storage Layers
- GPU Acceleration
- Encryption
- Compression
- Indexing
- Language Bindings
- WAL Recovery
- Beginner Tutorial
- User-Defined Functions (UDF)
- EventHook
- Job Scheduler
- SQL Triggers
- Query Plan Cache
- Stored Procedures
- Parallel Query
- Schema Versioning
- Feature Flags
- Database API
- SQL API
- Transaction API
- Roadmap
- License