初探可编程网关 Pipy

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"有幸参加了 ","attrs":{}},{"type":"link","attrs":{"href":"https://flomesh.cn/","title":"","type":null},"content":[{"type":"text","text":"Flomesh","attrs":{}}]},{"type":"text","text":" 组织的workshop，了解了他们的 Pipy 网络代理，以及围绕 Pipy 构建起来的生态。Pipy 在生态中，不止是代理的角色，还是 Flomesh 服务网格​中的数据平面。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"整理一下，做个记录，顺便瞄一下 Pipy 的部分源码。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"介绍","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"下面是摘自 Github 上关于 Pipy 的介绍：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 是一个轻量级、高性能、高稳定、可编程的网络代理。Pipy 核心框架使用 C++ 开发，网络 IO 采用 ASIO 库。 Pipy 的可执行文件仅有 5M 左右，运行期的内存占用 10M 左右，因此 Pipy 非常适合做 Sidecar proxy。","attrs":{}}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 内置了自研的 pjs 作为脚本扩展，使得Pipy 可以用 JS 脚本根据特定需求快速定制逻辑与功能。","attrs":{}}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 采用了模块化、链式的处理架构，用顺序执行的模块来对网络数据块进行处理。这种简单的架构使得 Pipy 底层简单可靠，同时具备了动态编排流量的能力，兼顾了简单和灵活。通过使用 REUSE_PORT 的机制（主流 Linux 和 BSD 版本都支持该功能），Pipy 可以以多进程模式运行，使得 Pipy 不仅适用于 Sidecar 模式，也适用于大规模的流量处理场景。 在实践中，Pipy 独立部署的时候用作“软负载”，可以在低延迟的情况下，实现媲美硬件的负载均衡吞吐能力，同时具有灵活的扩展性。","attrs":{}}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ff/ff8cc5a06afbb1c4944628ba0a135c6a.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 的核心是消息流处理器：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/e4/e43786b5003adb6f0df5706729813c42.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 流量处理的流程：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/03/03fcb7fceabb399a74fd8c3a360434ef.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"核心概念","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"流（Stream）：Pipy","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"管道（Pipeline）","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"模块（Module）","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"会话（Session）","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"上下文（Context）","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"以下是个人浅见：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 使用 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"pjs","attrs":{}}],"attrs":{}},{"type":"text","text":" 引擎将 JavaScript格式的配置，解析成其抽象的 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"Configuration","attrs":{}}],"attrs":{}},{"type":"text","text":" 对象。每个 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"Configuration","attrs":{}}],"attrs":{}},{"type":"text","text":" 中包含了多个 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"Pipeline","attrs":{}}],"attrs":{}},{"type":"text","text":"，每个 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"Configuration","attrs":{}}],"attrs":{}},{"type":"text","text":" 中又会用到多个 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"Filter","attrs":{}}],"attrs":{}},{"type":"text","text":"。这些都属于 Pipy 的","attrs":{}},{"type":"text","marks":[{"type":"italic","attrs":{}}],"text":"静态","attrs":{}},{"type":"text","text":"配置部分。（后面会提到 Pipeline 的三种不同类型）","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/8b/8b5b343ea0cc9509213baea71d053897.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"而属于","attrs":{}},{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"运行时","attrs":{}},{"type":"text","text":"的就是流、会话和上下文了，在 Pipy 中，数据流是由对象（Pipy 的","attrs":{}},{"type":"text","marks":[{"type":"italic","attrs":{}}],"text":"抽象","attrs":{}},{"type":"text","text":"）组成的。而这些对象抵达 Pipy，被抽象成不同的事件。而事件触发不同的过滤器的执行。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我个人更喜欢将其核心理解为：对数据流的事件处理引擎。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"理解归理解，实践出真知。“大胆假设，小心求证！”","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"本地编译","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"从编译 Pipy 开始。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"环境准备","attrs":{}}]},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"#安装 nodejs\n$ nvm install lts/erbium \n#安装 cmake\n$ brew install cmake\n","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"编译 Pipy","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"从 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"https://github.com/flomesh-io/pipy.git","attrs":{}}],"attrs":{}},{"type":"text","text":" 克隆代码。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Pipy 的编译包括了两个部分，GUI 和 Pipy 本体。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"GUI 是 Pipy 提供的一个用于开发模式下进行配置的界面，首先编译Pipy GUI。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"# pipy root folder\n$ cd gui\n$ npm install\n$ npm run build\n","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"接着编译 Pipy 的本体","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"# pipy root folder\n$ mkdir build\n$ cd build\n$ cmake -DCMAKE_BUILD_TYPE=Release -DPIPY_GUI=ON ..\n$ make\n","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"完成后检查根目录下的 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"bin","attrs":{}}],"attrs":{}},{"type":"text","text":" 目录，可以看到 pipy 的可执行文件，大小只有 11M。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/f3/f34ba26311cbecb1383af29fad836aff.jpeg","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"$ bin/pipy --help\nUsage: pipy [options]