Generate Unique ID by Redis 🚀

Introduction

This project provides an efficient and scalable unique ID generation system using Redis. Unlike traditional UUID-based approaches, this method optimizes database partitioning, improves query performance, and ensures structured ID generation.

Why Not Use UUID? 🤔

UUIDs (UUID.randomUUID()) are common for unique ID generation, but they have significant drawbacks:

• 🚫 Poor Indexing – UUIDs are random, making database indexing inefficient.
• 📏 Larger Storage – UUIDs require 128 bits, whereas numeric IDs are more compact.
• ❌ No Sequential Order – UUIDs do not provide a time-based order, reducing performance in sequential-write databases.

This project offers a partition-friendly ID format that ensures uniqueness while optimizing database performance.

---

Key Features ✨

🔹 Optimized for Database Partitioning

• ID Format: [Prefix][Timestamp][Sequence] (e.g., ORDER_231025_000000123)
  • Prefix: Categorizes data (e.g., ORDER, USER).
  • Timestamp: Date in yyMMdd format for time-based partitioning.
  • Sequence: Incremental number ensuring uniqueness.
• Benefits:
  • Efficient time-based partitioning in databases.
  • Faster queries and optimized storage.
  • Avoids fragmentation issues common with UUIDs.

🔹 Semi-Centralized Architecture

• Local Service: Generates IDs from a pre-reserved batch, reducing Redis dependency.
• Redis Backend: Tracks the highest reserved ID for global uniqueness.
• Benefits:
  • Minimizes latency by reducing Redis calls.
  • Ensures consistency across distributed systems.

🔹 Race Condition Prevention

• Producer-Consumer Model:
  • Producer: Reserves a batch of IDs from Redis and pre-generates IDs.
  • Consumer: Retrieves IDs from a thread-safe queue.
• Benefits:
  • Prevents race conditions in multi-threaded environments.
  • Ensures smooth ID generation without conflicts.

---

How It Works 🛠️

1. ID Preallocation

   • The local service requests a range of IDs from Redis (e.g., 1000 - 1999).
   • Redis updates the highest reserved ID and returns the range.

2. Local ID Generation

   • The service generates IDs using the format [Prefix][Timestamp][Sequence].
   • A LinkedBlockingQueue stores IDs for fast retrieval.

3. Concurrency Handling

   • Multiple threads retrieve IDs safely from the queue.
   • When the queue is almost empty, the producer requests a new batch from Redis.

4. Redis Ensures Global Uniqueness

   • Redis maintains the highest allocated ID, preventing duplication.

---

Components 🧩

• IdGenerator – Generates IDs based on reserved ranges.
• RangeProducer – Requests new ID ranges from Redis.
• RedisRangeProducer – Implements Redis-based ID reservation via Lua script.
• RedisConfig – Configures Redis connection and script integration.
• RedisApplication – Main application for multi-threaded ID generation.

---

Setup & Usage 📦

Prerequisites

• Java 11+
• Redis Server
• Spring Boot 2.7.7

Configuration (application.properties)

keyPrefix=ORDER       # Prefix for IDs (e.g., ORDER, USER)
reserveCount=100      # Number of IDs reserved per batch
numberProducer=2      # Number of producer threads

Running the Application

1. Start Redis.
2. Run the application:

   ./gradlew bootRun

---

Why This Approach? 🌟

✅ UUID vs. Redis-Based ID Comparison

Feature	Redis-based ID	UUID
Partitioning	✅ Optimized (time-based)	❌ Poor (random)
Performance	✅ High (batch reserve)	❌ Low (random)
Storage Size	✅ Compact (20-30 chars)	❌ Large (36 chars)
Sequential Order	✅ Yes (incremental)	❌ No
Complexity	⚖️ Moderate	✅ Simple

🚀 Benefits

• Scalability – Each service instance generates IDs independently.
• Efficiency – Reduces Redis calls for better performance.
• Better Partitioning – Unlike UUIDs, structured IDs help database indexing.

---

Contributing 🤝

Contributions are welcome! To contribute:

1. Fork the repository.
2. Create a new branch:

   git checkout -b feature/your-feature

3. Make your changes and commit them.
4. Push to the branch.
5. Open a Pull Request.

---

License 📄

This project is licensed under the MIT License. See LICENSE for details.