Mohamed KEITA

CortexDB v3

Advanced queries and a complete developer ecosystem.

1. Overall Vision of V3

V3 transforms CortexDB into a feature-rich engine with advanced queries and a complete developer ecosystem. It is divided into two sub-versions:

  • V3.1: Query Layer (advanced queries, lazy TTL, simple transactions)
  • V3.2: Developer Experience (CLI, server, clients, active TTL)

Main objectives:

  • Make CortexDB appealing to developers
  • Add advanced query capabilities
  • Simplify integration into applications
  • Build an ecosystem around CortexDB

2. V3 Structure

V3.1 — Query Layer

Features:

  • Range queries
  • Prefix queries (range-based sugar)
  • Lazy TTL (expiration checked at read time)
  • Simple transactions (strictly scoped)
  • Mini-schema (optional)

V3.2 — Developer Experience

Features:

  • Full CLI
  • Local server (HTTP/REST)
  • Python, TypeScript (Node.js)
  • Active TTL (cleanup + observability)
  • Integrations (appendix)

3. V3 Non-Goals

  • No vector search (V4)
  • No replication (V5)
  • No distributed transactions (V7)
  • No full MVCC
  • No sharding (V7)

4. V3 Architecture Overview

┌───────────┐
│   Client  │
│ (Python/  │
│TypeScript)│     │
└─────┬─────┘
      │  range_scan, prefix_scan, etc.

┌───────────────────┐
│   Query Engine    │
│ (Range/Prefix)    │
└─────┬─────────────┘


┌───────────────────┐
│  Transaction      │
│  Manager          │
└─────┬─────────────┘


┌───────────────────┐
│  Storage Engine   │
│  (V2)             │
└───────────────────┘

5. V3 Objective

At the end of V3 (V3.1 + V3.2), CortexDB becomes:

  • Feature-rich: advanced queries available
  • Developer-friendly: CLI, server, and multi-language clients
  • Production-ready: transactions and TTL for real-world use cases
  • Complete ecosystem: local server and Python, TypeScript, and Go clients

This is the version that makes CortexDB attractive and accessible to developers through a complete ecosystem.