分布式架构

架构演变

单体架构:将业务的所有功能集中在一个项目中开发,打成一个包部署

特点:简单方便、高度耦合、扩展性差,适合小型项目。例如:学生管理系统

分布式架构_第1张图片

分布式架构:根据业务功能对系统进行拆分,每个业务模块作为独立项目开发,称为一个服务

特点:松耦合,扩展性好,但架构复杂,难度大。适合大型互联网项目,例如:京东、淘宝

分布式架构_第2张图片

微服务:一种良好的分布式架构方案

服务治理:

分布式架构要考虑的问题:

  • 服务拆分粒度如何?
  • 服务集群地址如何维护?
  • 服务之间如何实现远程调用?
  • 服务健康状态如何感知?

分布式架构_第3张图片

微服务是一种经过良好架构设计的分布式架构方案,微服务架构特征:

  • 单一职责:微服务拆分粒度更小,每一个服务都对应唯一的业务能力,做到单一职责,避免重复业务开发
  • 面向服务: 微服务对外暴露业务接口
  • 自治:团队独立、技术独立、数据独立、部署独立
  • 隔离性强:服务调用做好隔离,容错,降级,避免出现级联问题

分布式架构_第4张图片

微服务技术栈:

分布式架构_第5张图片

SpringCloud

SpringCloud 是目前国内使用最广泛的微服务框架,官网地址:Spring Cloud

SpringCloud 集成了各种微服务功能组件,并基于SpringBoot 实现了这些组件的自动装配,从而提供了良好的开箱即用体验

与SpringBoot的版本兼容关系:

分布式架构_第6张图片

服务拆分及远程调用

服务拆分注意事项:

  • 不同微服务,不要重复开发相同业务
  • 微服务数据独立,不要访问其他微服务的数据库
  • 微服务可以将自己的业务暴露为接口,供其他微服务调用

分布式架构_第7张图片

案例:

项目结构:

分布式架构_第8张图片

启动项目后成功查询数据:

分布式架构_第9张图片

远程调用案例

需求:根据订单id查询订单的同时把订单所属的用户信息一起返回

分布式架构_第10张图片

步骤1:注入RestTemplate

package cn.itcast.order;

import org.mybatis.spring.annotation.MapperScan;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.context.annotation.Bean;
import org.springframework.web.client.RestTemplate;

@MapperScan("cn.itcast.order.mapper")
@SpringBootApplication
public class OrderApplication {

    public static void main(String[] args) {
        SpringApplication.run(OrderApplication.class, args);
    }

    @Bean
    public RestTemplate restTemplate(){
        return new RestTemplate();
    }
}

步骤2:修改订单业务逻辑,执行远程调用查询用户信息

package cn.itcast.order.service;

import cn.itcast.order.mapper.OrderMapper;
import cn.itcast.order.pojo.Order;
import cn.itcast.order.pojo.User;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;
import org.springframework.web.client.RestTemplate;

@Service
public class OrderService {

    @Autowired
    private OrderMapper orderMapper;

    @Autowired
    private RestTemplate restTemplate;

    public Order queryOrderById(Long orderId) {
        // 1.查询订单
        Order order = orderMapper.findById(orderId);
        //2.查询用户信息
        String url = "http://localhost:8081/user/" + order.getUserId();
        User user = restTemplate.getForObject(url, User.class);
        order.setUser(user);
        // 4.返回
        return order;
    }
}

效果:

分布式架构_第11张图片

服务调用关系:
1.服务提供者:暴露接口给其他微服务调用
2.服务消费者:调用其他微服务提供的接口
3.提供者与消费者角色是相对的
4.一个服务可以同时是服务者和服务消费者

Eureka 注册中心

作用:

分布式架构_第12张图片

消费者如何获取服务提供者具体信息?

  • 服务提供者启动时向eureka 注册自己的信息
  • eureka 保存这些信息
  • 消费者根据服务名称向eureka 拉取提供者信息

如果有多个服务提供者,消费者该如何选择?

  • 服务消费者利用负载均衡算法,从服务列表中挑选一个

消费者如何感知服务提供者健康状态?

  • 服务者会每隔30s向EurekaServer 发送心跳请求,报告健康状态
  • eureka 会更新记录服务列表信息,心跳不正常会被剔除
  • 消费者就可以拉去到最新的信息

Eurake 服务搭建

1.新建maven 工程引入依赖:

   
            org.springframework.cloud
            spring-cloud-starter-netflix-eureka-server
        

2.启动类:

package com.chuangzhou;

import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.cloud.netflix.eureka.server.EnableEurekaServer;


@EnableEurekaServer
@SpringBootApplication
public class EurekaServerApplication {


    public static void main(String[] args) {
        SpringApplication.run(EurekaServerApplication.class,args);
    }
}

3.编写配置:

eurekaserver 本身也是一个服务,因此也会注册自己的服务

server:
  port: 10086 # 服务端口
spring:
  application:
    name: eurekaserverdemo # eureka 的服务名称
eureka:
  client:
    service-url:
      defaultZone: http://127.0.0.1:10086/eureka  #eureka 的地址信息

eurekaserver 自己本身的服务:

分布式架构_第13张图片

服务注册

需求:将user-service、order-service 注册到 eureka-server

步骤:
1.引入Eureka 客户端依赖:

        
            org.springframework.cloud
            spring-cloud-starter-netflix-eureka-client
        

2.加入配置:

spring:
  application:
    name: orderserivce # eureka 的服务名称

eureka:
  client:
    service-url:
      defaultZone: http://127.0.0.1:10086/eureka #eureka 的地址信息

注册服务成功:

分布式架构_第14张图片

我们可以将user-service 多次启动,模拟多实例部署,但为了避免端口冲突,需要修改端口设置:

分布式架构_第15张图片

修改后启动:

分布式架构_第16张图片

服务发现

在order-service 完成服务拉取

服务拉取是基于服务名称获取服务列表,然后在服务列表做负载均衡

1,修改调用方式

package cn.itcast.order.service;

import cn.itcast.order.mapper.OrderMapper;
import cn.itcast.order.pojo.Order;
import cn.itcast.order.pojo.User;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;
import org.springframework.web.client.RestTemplate;

@Service
public class OrderService {

    @Autowired
    private OrderMapper orderMapper;

    @Autowired
    private RestTemplate restTemplate;

    public Order queryOrderById(Long orderId) {
        // 1.查询订单
        Order order = orderMapper.findById(orderId);
        //2.查询用户信息
//        String url = "http://localhost:8081/user/" + order.getUserId();
        String url = "http://userservice/user/" + order.getUserId();
        User user = restTemplate.getForObject(url, User.class);
        order.setUser(user);
        // 4.返回
        return order;
    }
}

2.为获取RestTemplate的方法添加负载均衡注解

    @Bean
    @LoadBalanced
    public RestTemplate restTemplate(){
        return new RestTemplate();
    }

分布式架构_第17张图片

分布式架构_第18张图片

Ribbon 负载均衡

负载均衡流程:

分布式架构_第19张图片

负载均衡策略

分布式架构_第20张图片

为 order-service 指定全局负载均衡策略:

    
    @Bean
    public IRule randomRule(){
        return new RandomRule();
    }

根据调用不同的服务指定不同的负载均衡策略,在application.yml 中配置即可:

userservice:
  ribbon:
    NFLoadBalancerRuleClassName: com.netflix.loadbalancer.RandomRule # 负载均衡规则

Ribbon 饥饿加载

默认为懒加载,因此当第一次请求时才会创建LoadBalanceClient,导致接口响应慢,之后的请求就快些:

分布式架构_第21张图片

通过控制台的log,我们也可以看到 当第一次请求来的时候才会执行初始化:

分布式架构_第22张图片

如何修改为饥饿加载?

ribbon:
  eager-load:
    enabled: true # 开启饥饿加载
    clients: userservice  # 指定对userserive 这个服务饥饿加载

项目启动后就会加载DynamicServerListLoadBalancer:

分布式架构_第23张图片

你可能感兴趣的:(分布式,架构)