🛍️ Product Feed View System

This project aims to evaluate how much Kafka combined with Redis caching can improve the performance of a distributed product feed system, in both response time and scalability.

📐 Architecture Overview

The system is composed of two microservices:

• Service 1 (Product Feed Consumer):
  Exposes one [GET] endpoint that returns a product list (similar to a shop feed like Amazon).

• Service 2 (Product Storage Provider):
  Stores and serves product data.

Communication between services is tested using two approaches:

• Synchronous REST communication
• Asynchronous communication via Kafka

⚙️ Technologies Used

• Java + Spring Boot
• Kafka
• Redis
• Docker

🚀 Performance Benchmark

Below are comparisons between different configurations, focusing on response time improvements.

Scenario 1: 300 Products

❌ REST Communication Only

Best result: 30ms

✅ Kafka + Redis

~73% improvement in response time with Kafka and Redis enabled. Best result: 8ms

✅ Kafka + Redis (Returning Raw JSON String)

~80% improvement when avoiding payload deserialization (raw String returned instead of Java object). Best result: 6ms

---

Scenario 2: 2000 Products

❌ REST Communication Only

Best result: 83ms

✅ Kafka + Redis

~80% improvement with asynchronous communication and caching. Best result: 16ms

✅ Kafka + Redis (Returning Raw JSON String)

~81% improvement, thanks to reduced processing overhead. Best result: 15ms

---

📈 Conclusion

This experiment shows that combining Redis with Kafka in a microservices architecture can lead to up to 80% faster response times, especially when avoiding heavy payload transformations.