负载均衡算法完全指南：从轮询到一致性哈希

import Chart from ‘…/…/components/Chart.vue’

export const weightedDistData = { labels: [‘节点 A (weight=3)’, ‘节点 B (weight=1)’, ‘节点 C (weight=2)’], datasets: [{ label: ‘请求占比 %’, data: [50, 16.7, 33.3], backgroundColor: [ ‘rgba(0,212,255,0.75)’, ‘rgba(255,0,170,0.75)’, ‘rgba(57,255,20,0.75)’, ], borderColor: [ ‘rgba(0,212,255,1)’, ‘rgba(255,0,170,1)’, ‘rgba(57,255,20,1)’, ], borderWidth: 1, }] }

export const weightedDistOptions = { plugins: { legend: { display: false }, tooltip: { callbacks: { label: (ctx) => ${ctx.raw}% } } }, scales: { x: { ticks: { color: ‘#8888aa’, font: { family: ‘JetBrains Mono’, size: 10 } }, grid: { color: ‘rgba(30,30,48,0.8)’ }, border: { display: false }, }, y: { min: 0, max: 60, ticks: { color: ‘#8888aa’, font: { family: ‘JetBrains Mono’, size: 10 }, callback: (v) => v + ‘%’ }, grid: { color: ‘rgba(30,30,48,0.8)’ }, border: { display: false }, } } }

export const algoSupportData = { labels: [‘轮询’, ‘加权轮询’, ‘随机’, ‘IP Hash’, ‘URL Hash’, ‘一致性Hash’, ‘最少连接’, ‘故障重试’], datasets: [ { label: ‘Nginx’, data: [5, 5, 5, 5, 5, 5, 5, 5], backgroundColor: ‘rgba(0,212,255,0.7)’, borderColor: ‘rgba(0,212,255,1)’, borderWidth: 1, borderRadius: 3, }, { label: ‘Dubbo’, data: [5, 5, 5, 0, 0, 5, 5, 5], backgroundColor: ‘rgba(57,255,20,0.7)’, borderColor: ‘rgba(57,255,20,1)’, borderWidth: 1, borderRadius: 3, }, { label: ‘Spring Cloud LB’, data: [5, 2, 5, 0, 0, 0, 0, 5], backgroundColor: ‘rgba(180,0,255,0.7)’, borderColor: ‘rgba(180,0,255,1)’, borderWidth: 1, borderRadius: 3, }, ] }

export const algoSupportOptions = { plugins: { legend: { labels: { color: ‘#c8c8d8’, font: { family: ‘JetBrains Mono’, size: 11 } } }, tooltip: { callbacks: { label: (ctx) => { const labels = [‘✖️ 不支持’, ‘’, ‘’, ‘’, ‘⚠️ 需自定义’, ‘✅ 支持’] return ${ctx.dataset.label}: ${ctx.raw === 5 ? '✅ 支持' : ctx.raw === 2 ? '⚠️ 需自定义' : '❌ 不支持'} } } } }, scales: { x: { ticks: { color: ‘#c8c8d8’, font: { family: ‘JetBrains Mono’, size: 9 }, maxRotation: 30 }, grid: { display: false }, border: { display: false }, }, y: { display: false, min: 0, max: 6, } } }

什么是负载均衡

负载均衡（Load Balancing）是指将客户端请求按照某种策略分发到多个后端服务节点，目标是：

• 避免单点过载：防止某一节点 CPU/内存耗尽
• 提升整体吞吐：充分利用集群资源
• 容灾与高可用：节点故障时自动剔除，流量转移

负载均衡分为两大类：静态算法（不考虑节点实时状态）和动态算法（根据节点当前负载调度）。

静态负载均衡算法

1. 轮询（Round Robin）

最简单的算法：请求依次分发给每个节点，循环往复。

请求序列: 1  2  3  4  5  6
节点:     A  B  C  A  B  C

优点：实现简单，分布均匀（假设节点同质）
缺点：不考虑节点性能差异和请求处理时长

Nginx 配置：

upstream backend {
    server 192.168.1.10;
    server 192.168.1.11;
    server 192.168.1.12;
    # 默认就是轮询
}

2. 随机（Random）

每次请求随机选择一个节点。大量请求下，统计上接近均匀分布。

优点：无状态，实现最简
缺点：小请求量时分布不均

3. 加权轮询（Weighted Round Robin）

根据节点处理能力分配权重，权重越高分得请求越多。

节点 A: weight=3, 节点 B: weight=1, 节点 C: weight=2
请求分配: A A A B C C A A A B C C ...

适用场景：集群中机器配置不均等（高配机器分更多流量）

upstream backend {
    server 192.168.1.10 weight=3;
    server 192.168.1.11 weight=1;
    server 192.168.1.12 weight=2;
}

weight=3:1:2 的实际请求分配比例：

hash(client_ip) % server_count = 目标节点索引

upstream backend {
    ip_hash;
    server 192.168.1.10;
    server 192.168.1.11;
}

upstream backend {
    hash $request_uri consistent;
    server 192.168.1.10;
    server 192.168.1.11;
}

          0
    ┌─────┼─────┐
  A(90°)  │   C(270°)
          │
        B(180°)

key_hash=120° → 顺时针 → C(270°)
key_hash=200° → 顺时针 → C(270°)
key_hash=300° → 顺时针 → A(90° + 360°)

节点 A: 活跃连接 100  → 不选
节点 B: 活跃连接 20   → 选择
节点 C: 活跃连接 55   → 不选

upstream backend {
    least_conn;
    server 192.168.1.10;
    server 192.168.1.11;
}

upstream backend {
    server prod.example.com  weight=9;
    server canary.example.com weight=1;
}