Miaosha is a high-concurrency “flash sale” (秒杀) demo that explores how to design a backend capable of withstanding extreme, bursty traffic during limited-time events. It illustrates patterns like cache-first reads for hot data, request rate limiting, and asynchronous order processing to protect databases under load. The project demonstrates stock deduction strategies designed to avoid overselling, combining pre-decrement in cache with reliable persistence and message-driven confirmation. Security hardening is addressed with ideas such as dynamic request paths, anti-bot measures, and one-time tokens to reduce automated abuse. The code showcases a typical Spring Boot stack that cleanly separates controllers, services, and persistence, making bottlenecks and optimization points easy to identify. As a study case, it provides a practical blueprint for building resilient sale/booking endpoints where fairness and correctness matter as much as raw throughput.
Features
- Microservices architecture with modules like admin, order, RPC service, frontend v1/v2
- High concurrency support using Redis, RabbitMQ, and ZooKeeper for messaging and coordination
- Integration with Spring Boot, Spring MVC, Dubbo for modular service organization
- Use of maven wrapper (mvnw) to simplify build and dependency management
- Detailed documentation and structure for deployment and experimentation
- Rich educational resource for engineering interviews and system design practice