📦 Distributed Cache System

A Node.js–based distributed in-memory cache system designed to reduce database load and improve read performance under concurrent traffic.
The system supports sharding, TTL-based expiration, LRU eviction, replication, failover simulation, and is Dockerized and load tested.

---

🚀 Features

• Distributed cache nodes
• Gateway-based request routing
• Hash-based sharding
• TTL-based expiration
• LRU eviction
• Replication and failover simulation
• Docker & Docker Compose support
• Load testing using k6

---

🏗️ Architecture (Textual)

Client → Gateway → Cache Nodes → Database

• All client requests go through a Gateway
• The Gateway shards keys across cache nodes
• Cache nodes store data in memory
• On cache miss, data is fetched from DB and cached

---

🔁 Request Flow (GET)

1. Client sends GET request to Gateway
2. Gateway routes request to cache shard
3. Cache hit → return value
4. Cache miss → fetch from DB
5. Store in cache with TTL
6. Return response to client

---

🧠 Replication & Consistency Model

• Hash-based sharding assigns keys to primary nodes
• Writes are replicated to a secondary node
• Reads are served from the primary
• Replica takes over on primary failure
• Eventual consistency model

---

🛠️ Tech Stack

• Node.js, Express
• Docker, Docker Compose
• AWS EC2
• k6 (Load Testing)
• Jest (Testing)

---

📡 API Endpoints

Set Cache Value

curl -X POST http://localhost:3000/cache \
-H "Content-Type: application/json" \
-d '{"key":"user123","value":"Alice","ttl":60}'

Get Cache Value
curl http://localhost:3000/cache/user123

Delete Cache Value
curl -X DELETE http://localhost:3000/cache/user123

📈 Performance

~6,000 requests/sec

Avg latency ~45 ms

P95 latency ~120 ms

Stable under node failure

🐳 Run with Docker
docker-compose up --build

👤 Author

Sanskriti Tyagi
GitHub: https://github.com/sanskritityagi31