Rate Limiter

A rate limiter built from scratch in Node.js and Redis to understand how different rate limiting algorithms work internally.

This project implements multiple algorithms instead of relying on existing libraries. The goal was to learn how modern rate limiters work, understand their trade-offs, and explore concurrency issues that appear in distributed systems.

---

What I Built

• Fixed Window
• Sliding Window
• Token Bucket
• Leaky Bucket
• Express middleware
• Redis integration
• Custom key generators
• Standard rate limit headers
• Retry-After support
• Custom limit exceeded handler

---

Why I Built It

Most developers use rate limiting libraries without understanding how they work internally.

This project was built to explore:

• How different rate limiting algorithms work
• The trade-offs between different strategies
• Redis data structures used by each algorithm
• Concurrency and race conditions
• How Redis can be used for scalable rate limiting

---

Algorithms

Algorithm	Accuracy	Burst Handling	Redis Memory	Best For
Fixed Window	Medium	Allows bursts at window boundaries	Low	Internal APIs
Sliding Window	High	Better burst control	Medium	Public APIs
Token Bucket	High	Excellent	Low	Authentication & API Gateways
Leaky Bucket	High	Constant request flow	Medium	Queues & Traffic Shaping

---

Quick Example

RateLimiter({
    strategy: "SlidingWindow",
    limit: 100,
    window: 60000,
    redis,
    keyGenerator: req => req.ip
});

---

Configuration

Option	Description
limit	Maximum requests allowed
window	Time window (milliseconds)
strategy	Rate limiting algorithm
redis	Connected Redis client
keyGenerator	Generates unique identifier
headers	Enables rate limit headers
identifier	Rate limit policy name
message	Default error message
onLimitReached	Custom callback

---

Current Status

Completed

• ✅ Fixed Window
• ✅ Sliding Window
• ✅ Token Bucket
• ✅ Leaky Bucket
• ✅ Redis integration
• ✅ Express middleware
• ✅ Retry-After support
• ✅ Response headers
• ✅ High concurrency testing

In Progress

• ⏳ Atomic Redis Lua scripts
• ⏳ EVALSHA optimization

Planned

• Cluster testing
• Multi-node benchmarking
• TypeScript support
• Better documentation
• Performance benchmarks

---

What I Learned

While building this project I explored:

• Redis Strings
• Redis Sorted Sets
• Redis Lists
• Rate limiting algorithms
• Redis expiration
• Concurrency
• Race conditions
• Distributed system fundamentals

One interesting finding was that a naive JavaScript implementation of Sliding Window suffers from race conditions under heavy concurrency, which is why production implementations usually rely on Redis Lua scripts for atomic execution.

---