本書主要介紹用于指導前端性能優(yōu)化工作的通用優(yōu)化方法，從網(wǎng)絡、瀏覽器、構(gòu)建工具、跨端技術(shù)和CDN 等方面介紹不同技術(shù)、系統(tǒng)對性能的影響，同時幫助讀者了解如何有效優(yōu)化性能。本書從性能的度量、分析和實驗這三個方面開始介紹。首先介紹性能優(yōu)化的一些通用方法，然后將性能作為一個切面幫助讀者了解與前端技術(shù)棧和性能有關(guān)的知識。從這個切面觀察，這些系統(tǒng)的工作原理等知識被賦予了另外一層意義，通過這種聯(lián)系把工作原理真正運用到工作中，在性能優(yōu)化方面發(fā)揮重要作用。本書面向的讀者為具有一定經(jīng)驗的 Web 開發(fā)工程師，以及對前端開發(fā)或 Web 開發(fā)有一定了解的開發(fā)人員。同時，假定讀者能夠進行簡單的網(wǎng)頁開發(fā)，并且具備相關(guān)的基礎(chǔ)知識。
佘錦鑫，花名當軒。畢業(yè)于江南大學理學院，阿里巴巴前前端技術(shù)專家，曾負責Alibaba.com海外版性能優(yōu)化，講師、開源愛好者。現(xiàn)就職于網(wǎng)易云音樂，對Web性能、跨端、可視化搭建等領(lǐng)域有較深入的理解。
第 1 篇 從 Vite 起步
第 1 章 從實踐開始 ··························.2
1.1 Hello World ···························.2
現(xiàn)在開始 ··························.2
使用 DevTools ····················.4
第一個優(yōu)化 ·······················.6
1.2 現(xiàn)實開發(fā)的例子 ······················.7
設(shè)置開發(fā)環(huán)境 ·····················.7
Vite ································.8
vite build ··························.9
進一步優(yōu)化 ·····················.11
引入 antd ························.11
按需引入 ························.13
動態(tài) import ·····················.14
1.3 小結(jié) ································.15
第 2 篇 性能優(yōu)化方法論
第 2 章 度量 ································.18
2.1 科學的方法 ·························.19
從一個客戶反饋說起 ··········.19
不度量性能，就無法優(yōu)化
性能 ······························.19
真實的用戶端性能 ·············.20
2.2 初識 Performance API ··············.21
performance.now()方法 ·········.21
構(gòu)建首屏指標 ···················.23
2.3 均值、分位數(shù)和秒開率 ············.23
均值 ·····························.24
分位數(shù) ···························.25
秒開率 ···························.26
如何選擇合適的統(tǒng)計指標 ·····.26
2.4 度量首屏 ···························.27
FP ································.27
FCP ······························.27
FMP ·····························.28
如何度量 FMP ··················.28
選定并度量首屏 ················.30
2.5 度量流暢度 ·························.30
度量流暢度的指標 ·············.31
可視化工具 ·····················.31
用戶端度量 ·····················.32
2.6 Core Web Vitals ····················.34
LCP ······························.34
FID ·······························.38
CLS ······························.39
2.6 小結(jié) ·································.41
第 3 章 分析 ································.42
3.1 分析方法 ···························.43
確定目標 ························.43
收集數(shù)據(jù) ························.43
清洗數(shù)據(jù) ························.44
統(tǒng)計值分析 ·····················.44
時序分析 ························.45
維度分析 ························.46
相關(guān)性分析 ·····················.48
3.2 常用的過程指標 ····················.48
TTFB ····························.49
DOMReady 和 Load ············.50
3.3 Performance API 詳解 ··············.51
Navigation Timing API ·········.51
Peformance Entry API ··········.53
Resource Timing ················.54
Navigation Timing Level 2 ·····.55
Paint Timing ····················.56
User Timing ·····················.56
3.4 分階段性能分析 ···················.58
常用的指標 ·····················.58
其他值得分析的指標 ··········.59
3.5 小結(jié) ································.59
第 4 章 實驗 ································.60
4.1 優(yōu)化不是照搬軍規(guī) ·················.61
時代在發(fā)展 ·····················.61
優(yōu)化的木桶效應明顯 ··········.62
用戶環(huán)境差異大 ················.62
性能實驗 ························.62
4.2 用實驗驗證優(yōu)化 ···················.63
混沌問題 ························.64
設(shè)計實驗 ························.64
分桶 ·····························.65
上報和分析數(shù)據(jù) ················.68
A/B Test 背后的數(shù)學 ···········.68
結(jié)論不重要，重要的是方法 ··.69
4.3 用實驗改進優(yōu)化 ···················.69
建立模型 ························.69
實驗修正 ························.70
4.4 小結(jié) ································.71
第 5 章 工具 ································.72
5.1 DevTools ···························.73
Network 面板 ···················.73
Performance 面板 ···············.76
5.2 WebPageTest ·······················.81
發(fā)起測試 ························.82
報告 ······························.83
Waterfall 視圖 ··················.83
5.3 小結(jié) ································.87
第 3 篇 網(wǎng)絡協(xié)議與性能
第 6 章 TTFB 為什么這么長 ··············.90
6.1 TTFB 的合理值 ····················.91
精確定義 ························.92
RTT ······························.92
RTT 一般需要多久 ·············.93
TTFB 的構(gòu)成····················.93
實驗環(huán)境驗證 ···················.94
6.2 如何優(yōu)化 TTFB ····················.95
減少請求的傳輸量 ·············.96
減少服務器端的處理時間 ·····.96
減少 RTT ························.98
TTFB 的值越小越好嗎 ·········.98
6.3 小結(jié) ·································.99
第 7 章 建立連接為什么這么慢 ·········.100
7.1 建立連接應該耗時多久 ···········.101
TCP 協(xié)議 ·······················.101
建立連接需要多少個 RTT ····.101
抓包驗證 ·······················.102
7.2 如何優(yōu)化建立連接的耗時 ········.103
減少物理距離 ··················.103
preconnect ······················.103
復用連接 ·······················.103
域名收攏 ·······················.104
TCP Fast Open ·················.104
QUIC 和 HTTP/3 ··············.104
7.3 小結(jié) ·······························.105
第 8 章 Fetch 之前瀏覽器在干什么 ····.106
8.1 重定向 ·····························.107
HTML 重定向 ·················.109
有哪些重定向 ·················.109
8.2 瀏覽器打開耗時 ··················.112
初始化標簽頁的時間 ·········.112
unload 的耗時 ··················.112
8.3 如何優(yōu)化 beforeFetch 耗時 ·······.114
重定向邏輯前置 ···············.115
合并重定向 ····················.115
避免使用短鏈 ·················.116
使用 beforeFetch 度量和
分析 ····························.116
8.4 小結(jié) ·······························.117
第 9 章 HTTPS 協(xié)議比 HTTP 協(xié)議
更慢嗎 ····························.118
9.1 HTTPS 協(xié)議為什么安全 ··········.119
對稱加密和非對稱加密 ·······.119
SSL/TLS 的實現(xiàn) ···············.120
SSL/TLS 握手 ·················.122
TLS False Start ·················.124
TLS 1.3 ·························.124
9.2 HTTPS 協(xié)議如何吊銷證書 ········.125
CRL ·····························.125
OCSP ···························.126
OCSP Stapling ·················.126
瀏覽器支持的情況 ············.126
證書類型 ·······················.127
證書驗證機制對性能的影響 ···.129
9.3 HTTPS 協(xié)議更慢嗎 ···············.129
確保證書鏈完整 ···············.129
啟用 TLS 1.3 ···················.129
不濫用 EV 證書 ···············.130
開啟 OSCP Stapling ···········.130
9.4 小結(jié) ································.130
第 10 章 HTTP/2、HTTP/3 和性能 ·····.131
10.1 HTTP/2 和性能 ···················.131
連接復用為什么不生效 ······.131
頭部壓縮對我們有什么影響 ··.137
為什么沒有廣泛使用 Server
Push ···························.140
10.2 為什么還需要 HTTP/3 ···········.144
HTTP/2 存在什么問題 ·······.145
HTTP/3 如何解決問題 ·······.146
10.3 小結(jié) ······························.148
第 11 章 壓縮和緩存 ·····················.150
11.1 傳輸速度和壓縮速度如何兼得 ···.151
Content-Encoding ·············.151
gzip 壓縮和 br 壓縮 ··········.152
實時壓縮 ······················.152
離線壓縮 ······················.153
如何優(yōu)化傳輸性能 ···········.154
11.2 HTTP 緩存什么時候會失效 ·····.154
緩存不僅僅是瀏覽器的
事情 ···························.154
緩存 Header ···················.154
11.3 小結(jié) ······························.157
第 4 篇 瀏覽器與性能
第 12 章 瀏覽器和性能 ···················.160
12.1 第一次渲染時都發(fā)生了什么 ····.161
最小的渲染路徑 ··············.162
盡快返回 HTML ··············.167
減少資源的阻塞 ··············.167
12.2 為什么 DOM 操作很慢 ··········.168
幀 ··································.168
重排 ·······························.169
重繪 ·······························.170
訪問 DOM 屬性 ················.170
如何優(yōu)化 DOM 操作 ··········.171
12.3 小結(jié) ······························.172
第 13 章 異步任務和性能 ················.173
13.1 事件循環(huán)機制 ····················.174
為什么要有事件循環(huán) ········.174
多線程阻塞模型 ··············.174
事件循環(huán) ······················.175
13.2 宏任務和微任務 ·················.176
13.3 Promise 的 polyfill 性能 ··········.178
如何正確實現(xiàn) Promise ·······.178
13.4 requestAnimationFrame ··········.180
13.5 小結(jié) ······························.181
第 14 章 內(nèi)存為什么會影響性能 ········.182
14.1 內(nèi)存 ······························.182
內(nèi)存管理 ·························.183
14.2 內(nèi)存泄漏 ·························.188
內(nèi)存泄漏和性能 ··············.188
常見的導致內(nèi)存泄漏的原因 ··.188
內(nèi)存泄漏問題的診斷工具 ···.189
14.3 小結(jié) ······························.191
第 15 章 使用 ServiceWorker 改善
性能 ······························.193
15.1 ServiceWorker 概述 ··············.194
AppCache······················.194
ServiceWorker·················.195
ServiceWorker 能做什么 ·····.195
15.2 使用 ServiceWorker 進行緩存 ···.196
Cache API ·····················.196
IndexDB ·······················.201
控制緩存的 Cache Key ······.201
更加靈活的緩存更新策略 ···.203
15.3 API 提前加載 ····················.204
15.4 ServiceWorker 冷啟動 ···········.205
開啟 Navigation Preload ·····.206
消費 Navigation Preload ·····.206
15.5 小結(jié) ······························.207
第 16 章 字體對性能的影響 ·············.208
16.1 字體導致的布局偏移 ············.208
如何定位布局偏移 ···········.208
16.2 如何避免字體帶來的布局偏移 ···.210
如何盡快加載字體 ···········.211
字體文件的格式 ··············.211
字體的加載 ···················.212
預加載字體 ···················.213
裁剪字體的大小 ··············.214
16.3 小結(jié) ······························.214
第 5 篇 前端工程與性能
第 17 章 構(gòu)建工具和性能 ················.218
17.1 為什么需要打包 ·················.219
CommonJS ····················.220
AMD ···························.220
CMD ···························.221
異步模塊加載器 ··············.222
依賴加載優(yōu)化 ················.223
模塊打包器 ···················.224
ES Module ····················.225
17.2 構(gòu)建工具可以做什么 ············.226
構(gòu)建工具和構(gòu)建優(yōu)化 ········.227
為什么要優(yōu)化打包體積 ·····.227
Bundle 分析 ···················.228
Tree Shaking ··················.229
Scope Hoisting ················.231
Code Splitting ·················.233
代碼壓縮 ······················.234
Vite 和 Bundleless ············.237
17.3 小結(jié) ······························.237
第 18 章 服務器端渲染和性能 ··········.239
18.1 SSR 和同構(gòu) ······················.241
18.2 SSR 的性能優(yōu)化 ·················.241
緩存 ·······························.242
數(shù)據(jù)預取 ·························.245
按需渲染 ·························.245
流式渲染 ·························.246
18.3 小結(jié) ······························.246
第 6 篇 跨端技術(shù)與 CDN
第 19 章 WebView 和性能 ···············.248
19.1 WebView 和 Native 的區(qū)別 ······.249
LayoutInflater ·················.249
加載 XML 的具體過程 ······.250
Measure ························.250
Layout ·························.251
Paint ···························.252
Surface ·························.253
SurfaceFlinger ·················.253
差異 ····························.253
19.2 WebView 的通信成本 ···········.254
JavaScript 調(diào)用 Native ········.254
Native 調(diào)用 JavaScript ········.258
雙向通信 ······················.258
通信對性能的影響 ···········.259
減少通信數(shù)據(jù)量 ··············.259
避免頻繁通信 ·················.259
19.3 React Native 的懶加載有何
不同 ······························.260
Web 實現(xiàn) ·····················.260
基于滾動容器的懶加載 ·····.260
基于位置獲取的懶加載 ·····.262
虛擬列表 ······················.263
19.4 React Native 如何減小打包
體積 ······························.265
Metro ··························.265
度量 ···························.266
分析 ···························.266
手動 Tree Shaking ············.267
利用 Babel 插件進行優(yōu)化 ···.269
體積和性能的關(guān)系 ···········.271
19.5 API 并行請求 ····················.271
發(fā)起請求 ·························.272
請求攔截 ·························.273
一致性檢驗 ······················.274
命中率分析 ······················.274
19.6 小結(jié) ······························.274
第 20 章 CDN 和性能 ····················.275
20.1 什么是 CDN ······················.275
解析 ·······························.276
邊緣節(jié)點 ·························.276
回源 ·······························.277
緩存策略 ·························.277
20.2 如何提升緩存命中率 ············.278
如何在端側(cè)統(tǒng)計緩存命中的
情況 ····························.278
減少緩存分裂 ·················.279
緩存忽略動態(tài)參數(shù) ···········.279
歸一化 Vary Header ··········.280
長效緩存 ······················.280
20.3 動態(tài)加速 ·························.281
海外加速 ·························.282
連接復用 ·························.282
客戶端連接復用 ················.282
HTTPS 優(yōu)化 ·····················.283
動靜分離 ·························.283
壓縮 ·······························.284
什么場景適合使用動態(tài)加速 ··.284
20.4 自動 polyfill ······················.284
什么是 polyfill ················.284
Polyfill.io ······················.285
實現(xiàn)原理 ······················.287
20.5 邊緣計算和性能 ·················.288
CDN 的可編程功能 ··········.288
Hello World ···················.289
自定義 Cache Key ············.289
前置重定向 ···················.290
流式渲染 ······················.290
20.6 小結(jié) ······························.291
新書資訊