# EloqData
> SSD-first, Memory-Speed Persistence Database Architecture.

## Project Overview
EloqData is building the next generation of high-performance databases designed for the modern hardware landscape (NVMe SSDs) rather than traditional DRAM-heavy models. Our architecture focuses on extreme low tail latency (P99.99 < 6ms) and massive concurrency.

## Core Technology
- **EloqStore**: An open-source, SSD-first storage engine.
- **Async I/O**: Deep integration with Linux io_uring for high-throughput disk operations.
- **Concurrency**: Built using C++20 coroutines to handle millions of concurrent requests.
- **APIs**: Provides a multi-model layer including:
    - **EloqKV**: Redis-compatible API.
    - **EloqDoc**: MongoDB-compatible API.
    - Built on top of ScyllaDB for horizontal scalability.

## Key Performance Targets
- Optimized for "Memory-Speed Persistence" (using SSDs as primary storage while maintaining DRAM-like speeds).
- Targeted at AI workloads, real-time analytics, and high-scale metadata management.
- Designed to bypass global DRAM shortages and high infrastructure costs.

## Documentation & Resources
- Main Site: https://www.eloqdata.com
- GitHub: https://github.com/EloqData
- Technical Blog: https://www.eloqdata.com/blog

## Team
- Co-founded by Lintao Zhang (CEO), Hubert Zhang (CTO) and Jeff Chen (Chief Architect).
- Engineering roots in GreenplumDB and Pivotal database internals.