Microservice Architecture: Technology Selection, Service Governance, API Gateways, and Configuration Centers

1. Introduction When rebuilding a monolithic application into a microservice architecture, developers must choose among many technical frameworks; large companies often build their own, while small‑to‑medium enterprises rely on open‑source solutions.

2. Selection Principles The author proposes several guidelines: introduce a technology only when there is a concrete need, prefer the most familiar and widely used tools, choose components with strong community support (e.g., high GitHub stars, active issue resolution), consider business size and project scale, and validate new technologies before large‑scale adoption.

3. Technology Options A comparative table (omitted) lists common microservice frameworks. The article then examines each in detail.

Dubbo Dubbo is a high‑performance, lightweight Java RPC framework offering remote method invocation, fault tolerance, load balancing, and service registration/discovery. Its logical architecture includes Provider, Consumer, Registry, Monitor, and Container. While strong in RPC efficiency and community adoption in China, Dubbo lacks many higher‑level features such as distributed configuration, tracing, and batch‑task support.

Spring Cloud Spring Cloud is the most popular microservice stack built on Spring Boot, providing a comprehensive suite of modules for service registration, dynamic routing, load balancing, configuration management, message bus, circuit breaking, distributed tracing, and big‑data integration. It integrates tightly with Spring Boot and is considered the default choice for many enterprises.

4. API Gateways

Zuul As a core component of Spring Cloud, Zuul acts as an API gateway with dynamic routing, authentication, monitoring, load‑balancing, and other filter‑based features. Filters are implemented in Groovy and communicate via a ThreadLocal‑based RequestContext.

Traefik Traefik is an HTTP reverse‑proxy and load‑balancer written in Go. It automatically discovers services from back‑ends such as Docker, Kubernetes, Consul, etc., supports hot‑reloading of configuration, SSL, HTTP/2, and provides a clean AngularJS UI.

Kong Kong runs on top of OpenResty (Nginx + Lua) and offers a plugin‑based API gateway with dynamic routing, load balancing, authentication, rate limiting, SSL, health checks, and real‑time monitoring. Its plugin architecture makes it highly extensible.

5. Service Registration & Discovery Traditional DNS+LVS+Nginx solutions gave way to dynamic registries such as Zookeeper, Eureka, Consul, and etcd, which provide real‑time service lookup, health checking, and load‑balancing capabilities.

6. Configuration Centers

Spring Cloud Config Provides centralized configuration stored in Git, SVN, or databases, with a Config Server, optional Bus for change propagation, and supports multiple environments.

Apollo Ctrip’s distributed configuration platform with MySQL storage, Config Service, Admin Service, and a web portal. It offers real‑time push, permission control, and gray‑release workflows.

Nacos An easy‑to‑use dynamic service discovery and configuration platform for cloud‑native applications. It supports DNS‑based and RPC‑based discovery, dynamic configuration, and integrates with Spring Cloud.

7. Comparative Summary Spring Cloud offers the richest module ecosystem, Dubbo excels in pure RPC performance, Zuul and Traefik are strong API‑gateway choices, while Kong provides the highest performance and plugin flexibility. For configuration management, Apollo and Nacos have broader ecosystem support than Spring Cloud Config, with Apollo offering more comprehensive features and Nacos being simpler to operate.

References (list of URLs omitted for brevity).