Evolution of a Java Web E‑commerce System: From Single‑Server to Microservices and Distributed Architecture

We use a Java Web example to build a simple e‑commerce system and demonstrate how it can evolve through multiple stages.

Stage 1 – Single‑machine deployment : All components (Tomcat/Jetty, JSP/Servlet, Spring, MyBatis, MySQL, etc.) run on one server, forming a minimal system.

Stage 2 – Separate application and database servers : To handle increasing traffic, the web server and database server are split, improving load capacity and fault tolerance.

Stage 3 – Application server clustering : Multiple application servers are added behind a load balancer (e.g., Keepalived + ipvsadm). Various load‑balancing methods (HTTP redirect, DNS, reverse proxy, IP‑layer, data‑link layer) are introduced.

Stage 4 – Database read/write separation : Master‑slave replication or middleware (e.g., MyCat) is used to separate read and write traffic, addressing database load.

Stage 5 – Search engine for fuzzy queries : To improve product search, an inverted‑index search engine is added, reducing the cost of LIKE queries.

Stage 6 – Caching layer : Application‑level (Guava, Memcached) and database‑level (Redis) caches are introduced, along with page caching (localStorage, static pages) to relieve database pressure.

Stage 7 – Database vertical and horizontal sharding : Vertical sharding separates business domains (users, products, orders) into different databases; horizontal sharding splits large tables across multiple databases, with middleware handling routing, primary‑key generation, and pagination.

Stage 8 – Microservice decomposition : The monolithic application is split into smaller services (e.g., user‑service, product‑service, order‑service) to avoid code bloat and improve maintainability.

Stage 9 – Messaging middleware : Introducing a language‑agnostic message bus (e.g., Dubbo with Zookeeper) enables reliable inter‑service communication, load‑balancing transparency, and operational monitoring.

The final architecture combines load balancers, clustered application servers, read/write‑separated databases, search engines, caching clusters, sharded databases, microservices, and a messaging layer, providing a scalable and resilient solution for large‑scale web applications.