十万同时在线用户，需要多少内存？——Newbe.Claptrap 框架水平扩展实验

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Newbe.Claptrap 项目是笔者正在构建以"},{"type":"codeinline","content":[{"type":"text","text":"反应式"}]},{"type":"text","text":"、"},{"type":"codeinline","content":[{"type":"text","text":"Actor模式"}]},{"type":"text","text":"和"},{"type":"codeinline","content":[{"type":"text","text":"事件溯源"}]},{"type":"text","text":"为理论基础的一套服务端开发框架。本篇我们将来了解一下框架在水平扩展方面的能力。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"前情提要"}]},{"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":"时隔许久，今日我们再次见面。首先介绍一下过往的项目情况："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/Overview-Of-Newbe-Claptrap/","title":null},"content":[{"type":"text","text":"第一次接触本框架的读者，可以先点击此处阅读本框架相关的基础理论和工作原理。"}]}]},{"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":"日前，我们也编写了一些预热文章和工具，读者可以通过以下链接进行了解："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/Reactive-In-Server-1/","title":null},"content":[{"type":"text","text":"谈反应式编程在服务端中的应用，数据库操作优化，从 20 秒到 0.5 秒"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Others/Full-Speed-To-Download-Netcore-Docker-Images/","title":null},"content":[{"type":"text","text":"docker-mcr 助您全速下载 dotnet 镜像"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/Newbe-Claptrap-Weekly-1/","title":null},"content":[{"type":"text","text":"Newbe.Claptrap 项目周报 1 - 还没轮影，先用轮跑"}]}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"今日主题"}]},{"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":"今天，我们来做一套实验预演，来验证 Newbe.Claptrap 框架，如何通过水平扩展的形式来适应逐渐增长的同时在线用户数。"}]},{"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":"由于此次实验涉及的内容很多，因此笔者将内容进行了归类，读者可以按照自己的兴趣阅读相关的章节："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#业务需求说明","title":null},"content":[{"type":"text","text":"业务需求说明"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#调用时序关系","title":null},"content":[{"type":"text","text":"调用时序关系"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#物理结构设计","title":null},"content":[{"type":"text","text":"物理结构设计"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#实际测试数据","title":null},"content":[{"type":"text","text":"实际测试数据"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#源码构建说明","title":null},"content":[{"type":"text","text":"源码构建说明"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/#常见问题解答","title":null},"content":[{"type":"text","text":"常见问题解答"}]}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"业务需求说明"}]},{"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":"先看看今天要实现的业务场景："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"用户通过 API 登录后生成一个 JWT token"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"用户调用 API 时验证 JWT token 的有效性"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"没有使用常规的 JWS 公私钥方式进行 JWT token 颁发，而是为每个用户单独使用 secret 进行哈希验证"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"验证看不同的在线用户需要消耗的内存情况"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"用户登录到生成 token 所消耗时间不得超过 200 ms"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"tokn 的验证耗时不得超过 10 ms"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"吹牛先打草稿"}]},{"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":"笔者没有搜索到于 “在线用户数” 直接相关的理论定义，因此，为了避免各位的理解存在差异。笔者先按照自己的理解来点明：在线用户数到底意味着什么样的技术要求？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"未在线用户若上线，不应该受到已在线用户数的影响"}]},{"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":"如果一个用户登录上线需要消耗 100 ms。那么不论当前在线的用户数是十人还是百万人。这个登录上线所消耗的时间都不会明显的超过 100 ms。"}]},{"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":"当然，有限的物理硬件肯定会使得，当在线用户数超过一个阈值（例如两百万）时，新用户登录上线会变慢甚至出错。"}]},{"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":"但是，增加物理机器就能提高这个阈值，我们就可以认为水平扩展设计是成功的。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"对于任意一个已在线用户，得到的系统性能反馈应当相同"}]},{"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":"例如已在线的用户查询自己的订单详情，需要消耗 100 ms。那么当前任何一个用户进行订单查询的平均消耗都应该稳定在 100 ms。"}]},{"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":"当然，这里需要排除类似于 “抢购” 这种高集中性能问题。此处主要还是讨论日常稳定的容量增加。（我们以后会另外讨论 “抢购” 这种问题）"}]},{"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":"具体一点可以这样理解。假设我们做的是一个云笔记产品。"}]},{"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":"那么，如果增加物理机器就能增加同时使用云笔记产品的用户数，而且不牺牲任何一个用户的性能体验，我们就认为水平扩展设计是成功的。"}]},{"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":"在此次的实验中，若用户已经登录，则验证 JWT 有效性的时长大约为 0.5 ms。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"调用时序关系"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/c5/c5275fe97967395c2efa4398c0f4503c.png","alt":"Timing diagram","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"简要说明："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"numberedlist","attrs":{"start":null,"normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"客户端发起登录请求将会逐层传达到 UserGrain 中"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"text","text":"UserGrain 将会在内部激活一个 Claptrap 来进行维持 UserGrain 中的状态数据。包括用户名、密码和用于 JWT 签名的 Secret。"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":3,"align":null,"origin":null},"content":[{"type":"text","text":"随后的生成 JWT 生成和验证都将直接使用 UserGrain 中的数据。由于 UserGrain 中的数据是在一段时间内是 “缓存” 在内存中的。所以之后的 JWT 生成和验证将非常快速。实测约为 0.5 ms。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"物理结构设计"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/2d/2de360a7777693f9ab2768abc86b30c4.png","alt":"物理结构设计","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如上图所示，便是此次进行测试的物理组件："}]},{"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":"名称说明WebAPI公开给外部调用 WebAPI 接口。提供登录和验证 token 的接口。Orleans Cluster托管 Grain 的核心进程.Orleans Gateway于 Orleans Cluster 基本相同，但是 WebAPI 只能与 Gateway 进行通信Orleans Dashboard于 Orleans Gateway 基本相同，但增加了 Dashboard 的展示，以查看整个 Orleans 集群的情况Consul用于 Orleans 集群的集群发现和维护Claptrap DB用于保存 Newbe.Claptrap 框架的事件和状态数据Influx DB & Grafana用于监控 Newbe.Claptrap 相关的性能指标数据"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"此次实验的 Orleans 集群节点的数量实际上是 Cluster + Gateway + Dashboard 的总数。以上的划分实际上是由于功能设定的不同而进行的区分。"}]},{"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":"此次测试 “水平扩展” 特性的物理节点主要是 Orleans Cluster 和 Orleans Gateway 两个部分。将会分别测试以下这些情况的内存使用情况。"}]},{"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":"Orleans DashboardOrleans GatewayOrleans Cluster100111135"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"此次实验采用的是 Windows Docker Desktop 结合 WSL 2 进行的部署测试。"}]},{"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":"以上的物理结构实际上是按照最为此次实验最为复杂的情况设计的。实际上，如果业务场景足够简单，该物理结构可以进行裁剪。详细可以查看下文 “常见问题解答” 中的说明。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"实际测试数据"}]},{"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，分别对不同的集群规模和用户数量进行测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"0 Gateway 0 Cluster"}]},{"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":"默认情况下，刚刚启动 Dashboard 节点时，通过 portainer 可以查看 container 占用的内存约为 200 MB 左右，如下图所示："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/fa/fa60699c21cee4d1e665982081c1cfc8.png","alt":"初始内存占用","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通过测试控制台，向 WebAPI 发出 30,000 次请求。每批 100 个请求，分批发送。"}]},{"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":"经过约两分钟的等待后，再次查看内存情况，约为 9.2 GB，如下图所示："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/1c/1cb116edae9f332a575c5498f04106c7.png","alt":"三万用户","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"因此，我们简单的估算每个在线用户需要消耗的内存情况约为 （9.2*1024-200）/30000 = 0.3 MB。"}]},{"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":"另外，可以查看一些辅助数据："}]},{"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":"CPU 使用情况"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ff/ffbe2da3475c36bb0c187fbf4dc6f1ef.png","alt":"CPU使用情况","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"网络吞吐量"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/f6/f67f7ff3041ae67f974d4fe91124238c.png","alt":"网络吞吐量","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Orleans Dashboard 情况。左上角的 TOTAL ACTIVATIONS 中 30,000 即表示当前内存中存在的 UserGrain 数量，另外的 3 个为 Dashboard 使用的 Grain。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/20/207d5ae371e861abcf0debe80b1dc90b.png","alt":"Orleans Dashboard 情况","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Grafana 中查看 Newbe.Claptrap 的事件平均处理时长约为 100-600 ms。此次测试的主要是内存情况，处理时长的采集时间为 30s 一次，因此样本数并不多。关于处理时长我们将在后续的文章中进行详细测试。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/8d/8dbb09fc13fc148c7d04a71806d84bee.png","alt":"时间平均处理时长","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Grafana 中查看 Newbe.Claptrap 的事件的保存花费的平均时长约为 50-200 ms。事件的保存时长是事件处理的主要部分。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/93/93813be7e521efb96a3fc1184c4e45ba.png","alt":"三万用户","title":null,"style":null,"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}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Grafana 中查看 Newbe.Claptrap 的事件已处理总数。一种登录了三万次，因此事件总数也是三万。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/d3/d36e5f48ab9df5e7ed45049c1cad4548.png","alt":"事件处理的总数","title":null,"style":null,"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}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"1 Gateway 1 Cluster"}]},{"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":"接下来，我们测试额外增加两个节点进行测试。"}]},{"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":"还是再提一下，Orleans 集群节点的数量实际上是 Cluster + Gateway + Dashboard 的总数。因此，对比上一个测试，该测试的节点数为 3。"}]},{"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":"测试得到的内存使用情况如下："}]},{"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":"用户数节点平均内存内存总占用100001.8 GB1.8*3 = 5.4 GB200003.3 GB3.3*3 = 9.9 GB300004.9 GB4.9*3 = 14.7 GB"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那么，以三万用户为例，平均每个用户占用的内存约为 （14.7*1024-200*3）/30000 = 0.48 MB"}]},{"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":"为什么节点数增加了，平均消耗内存上升了呢？笔者推测，没有进行过验证：节点增加，实际上节点之间的通讯还需要消耗额外的内存，因此平均来说有所增加。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"3 Gateway 5 Cluster"}]},{"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":"我们再次增加节点。总结点数为 1 （dashboard） + 3 （cluster） + 5 (gateway) = 9 节点"}]},{"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":"测试得到的内存使用情况如下："}]},{"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":"用户数节点平均内存内存总占用200001.6 GB3.3*9 = 14.4 GB300002 GB4.9*9 = 18 GB"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那么，以三万用户为例，平均每个用户占用的内存约为 （18*1024-200*9）/30000 = 0.55 MB"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"十万用户究竟要多少内存？"}]},{"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":"以上所有的测试都是以三万为用户数进行的测试，这是一个特殊的数字。因为继续增加用户数的话，内存将会超出测试机的内存余量。（求赞助两条 16G）"}]},{"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":"如果继续增加用户数，将会开始使用操作系统的虚拟内存。虽然可以运行，但是运行效率会降低。原来登录可能只需要 100 ms。使用到虚拟内存的用户则需要 2 s。"}]},{"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":"因此，速度降低的情况下，在验证需要多少内存意义可能不大。"}]},{"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":"但是，这不意味着不能够继续登录，以下便是 1+1+1 的情况下，十万用户全部登录后的情况。(有十万用户同时在线，加点内存吧，不差钱了。)"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/a1/a1b79458d84072f24f7f73a9fb9a6de4.png","alt":"十万用户","title":null,"style":null,"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}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"源码构建说明"}]},{"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":"此次测试的代码均可以在文末的样例代码库中找到。为了方便读者自行实验，主要采用的是 docker-compose 进行构建和部署。"}]},{"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":"因此对于测试机的唯一环境需求就是要正确的安装好 Docker Desktop 。"}]},{"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":"可以从以下任一地址获取最新的样例代码："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://github.com/newbe36524/Newbe.Claptrap.Examples","title":null},"content":[{"type":"text","text":"https://github.com/newbe36524/Newbe.Claptrap.Examples"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://gitee.com/yks/Newbe.Claptrap.Examples","title":null},"content":[{"type":"text","text":"https://gitee.com/yks/Newbe.Claptrap.Examples"}]}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"快速启动"}]},{"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":"使用控制台进入 "},{"type":"codeinline","content":[{"type":"text","text":"src/Newbe.Claptrap.Auth/LocalCluster"}]},{"type":"text","text":" 文件夹。运行以下命令便可以在本地启动所有的组件："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"docker-compose up -d"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"途中需要拉取一些托管于 Dockerhub 上的公共镜像，请确保本地已经正确配置了相关的加速器，以便您可以快速构建。"},{"type":"link","attrs":{"href":"https://www.runoob.com/docker/docker-mirror-acceleration.html","title":null},"content":[{"type":"text","text":"可以参看这篇文档进行设置"}]}]},{"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":"成功启动之后可以通过 "},{"type":"codeinline","content":[{"type":"text","text":"docker ps"}]},{"type":"text","text":" 查看到所有的组件。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"PS>docker ps\nCONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES\n66470e5393e2 registry.cn-hangzhou.aliyuncs.com/newbe36524/newbe-claptrap-auth-webapi \"dotnet Newbe.Claptr…\" 4 hours ago Up About an hour 0.0.0.0:10080->80/tcp localcluster_webapi_1\n3bbaf5538ab9 registry.cn-hangzhou.aliyuncs.com/newbe36524/newbe-claptrap-auth-backendserver \"dotnet Newbe.Claptr…\" 4 hours ago Up About an hour 80/tcp, 443/tcp, 0.0.0.0:19000->9000/tcp, 0.0.0.0:32785->11111/tcp, 0.0.0.0:32784->30000/tcp localcluster_dashboard_1\n3f60f51e4641 registry.cn-hangzhou.aliyuncs.com/newbe36524/newbe-claptrap-auth-backendserver \"dotnet Newbe.Claptr…\" 4 hours ago Up About an hour 80/tcp, 443/tcp, 9000/tcp, 0.0.0.0:32787->11111/tcp, 0.0.0.0:32786->30000/tcp localcluster_cluster_gateway_1\n7d516ada2b26 registry.cn-hangzhou.aliyuncs.com/newbe36524/newbe-claptrap-auth-backendserver \"dotnet Newbe.Claptr…\" 4 hours ago Up About an hour 80/tcp, 443/tcp, 9000/tcp, 30000/tcp, 0.0.0.0:32788->11111/tcp localcluster_cluster_core_1\nfc89fcd973f9 grafana/grafana \"/run.sh\" 4 hours ago Up 6 seconds 0.0.0.0:23000->3000/tcp localcluster_grafana_1\n1f10ed0eb25f postgres \"docker-entrypoint.s…\" 4 hours ago Up About an hour 0.0.0.0:32772->5432/tcp localcluster_claptrap_db_1\nd5d2bec74311 adminer \"entrypoint.sh docke…\" 4 hours ago Up About an hour 0.0.0.0:58080->8080/tcp localcluster_adminer_1\n4c4be69f2f41 bitnami/consul \"/opt/bitnami/script…\" 4 hours ago Up About an hour 8300-8301/tcp, 8500/tcp, 8301/udp, 8600/tcp, 8600/udp localcluster_consulnode3_1\n88811d3aa0d2 influxdb \"/entrypoint.sh infl…\" 4 hours ago Up 6 seconds 0.0.0.0:29086->8086/tcp localcluster_influxdb_1\nd31c73b62a47 bitnami/consul \"/opt/bitnami/script…\" 4 hours ago Up About an hour 8300-8301/tcp, 8500/tcp, 8301/udp, 8600/tcp, 8600/udp localcluster_consulnode2_1\n72d4273eba2c bitnami/consul \"/opt/bitnami/script…\" 4 hours ago Up About an hour 0.0.0.0:8300-8301->8300-8301/tcp, 0.0.0.0:8500->8500/tcp, 0.0.0.0:8301->8301/udp, 0.0.0.0:8600->8600/tcp, 0.0.0.0:8600->8600/udp localcluster_consulnode1_1"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"启动完成之后，便可以通过以下链接来查看相关的界面"}]},{"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":"地址说明"},{"type":"link","attrs":{"href":"http://localhost:19000/","title":null},"content":[{"type":"text","text":"http://localhost:19000"}]},{"type":"text","text":"Orleans Dashboard 查看 Orleans 集群中各节点的状态"},{"type":"link","attrs":{"href":"http://localhost:10080/","title":null},"content":[{"type":"text","text":"http://localhost:10080"}]},{"type":"text","text":"Web API 基地址，此次使用所测试的 API 基地址"},{"type":"link","attrs":{"href":"http://localhost:23000/","title":null},"content":[{"type":"text","text":"http://localhost:23000"}]},{"type":"text","text":"Grafana 地址，查看 Newbe.Claptrap 相关的性能指标情况"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"源码构建"}]},{"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":"使用控制台进入 "},{"type":"codeinline","content":[{"type":"text","text":"src/Newbe.Claptrap.Auth"}]},{"type":"text","text":" 文件夹。运行以下命令便可以在本地完成代码的构建："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"./LocalCluster/pullimage.cmd\ndocker-compose build"}]},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"pullimage.cmd 使用了笔者编写的 docker-mcr 加速器功能。"},{"type":"link","attrs":{"href":"https://www.newbe.pro/Others/Full-Speed-To-Download-Netcore-Docker-Images/","title":null},"content":[{"type":"text","text":"您可以通过该文档来了解其工作原理"}]}]}]},{"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":"等待构建完毕之后，本地便生成好了相关的镜像。接下来便可以初次尝试在本地启动应用："}]},{"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":"使用控制台进入 "},{"type":"codeinline","content":[{"type":"text","text":"src/Newbe.Claptrap.Auth/LocalCluster"}]},{"type":"text","text":" 文件夹。运行以下命令便可以启动相关的容器:"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"shell"},"content":[{"type":"text","text":"docker-compose up -d"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"常见问题解答"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"文中为何没有说明代码和配置的细节？"}]},{"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":"本文主要为读者展示该方案的实验可行性，具体应该如何应用 Newbe.Claptrap 框架编写代码，并非本文的主旨，因此没有提及。"}]},{"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":"当然，另外一点就是目前框架没有最终定版，所有内容都有可能发生变化，讲解代码细节意义不大。"}]},{"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":"但可以提前说明的是：编写非常简单，由于本样例的业务需求非常简单，因此代码内容也不多。全部都可以在示例仓库中找到。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"用 Redis 存储 Token 也可以实现上面的需求，为什么要选择这个框架？"}]},{"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":"目前来说，笔者没有十足的理由说服读者必须使用哪种方案，此处也只是提供一种可行方案，至于实际应该选择哪种方案，应该有读者自己来考量，毕竟工具是否趁手还是需要试试才知道。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"如果是最多 100 个在线用户，那怎么裁剪系统？"}]},{"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":"必要的组件只有 Orleans Dashboard 、 WebAPI 和 Claptrap Db。其他的组件全部都是非必要的。而且如果修改代码， Orleans Dashboard 和 WebAPI 是可以合并的。"}]},{"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":"所以最小规模就是一个进程加一个数据库。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"Grafana 为什么没有报表？"}]},{"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":"Grafana 首次启动之后需要手动的创建 DataSource 和导入 Dashboard."}]},{"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":"本实验相关的参数如下："}]},{"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":"DataSource"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"URL： "},{"type":"link","attrs":{"href":"http://influxdb:8086/","title":null},"content":[{"type":"text","text":"http://influxdb:8086"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Database： metricsdatabase"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"User： claptrap"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Password： claptrap"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://github.com/newbe36524/Newbe.Claptrap/blob/develop/src/Docker/Monitor/grafana/claptrap.json","title":null},"content":[{"type":"text","text":"点击此处获取 Dashboard 定义文件"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"测试机的物理配置是什么？"}]},{"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":"没有专门腾内存，未开始测试前已占用 16GB 内存。以下是测试机的身材数据（洋垃圾，3500 元左右）："}]},{"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":"处理器 英特尔 Xeon (至强) E5-2678 v3 @ 2.50GHz 12 核 24 线程"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"主板 HUANANZHI X99-AD3 GAMING (Wellsburg)"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"显卡 Nvidia GeForce GTX 750 Ti (2 GB / Nvidia)"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"内存 32 GB (三星 DDR3L 1600MHz) 2013 年产 高龄内存"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"主硬盘 金士顿 SA400S37240G (240 GB / 固态硬盘)"}]},{"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":"如果您有更好的物理配置，相信可以得出更加优秀的数据。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"即使是 0.3 MB 平均每用户的占用的我也觉得太高了"}]},{"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":"框架还在优化。未来会更好。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"最后但是最重要！"}]},{"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":"最近作者正在构建以"},{"type":"codeinline","content":[{"type":"text","text":"反应式"}]},{"type":"text","text":"、"},{"type":"codeinline","content":[{"type":"text","text":"Actor模式"}]},{"type":"text","text":"和"},{"type":"codeinline","content":[{"type":"text","text":"事件溯源"}]},{"type":"text","text":"为理论基础的一套服务端开发框架。希望为开发者提供能够便于开发出 “分布式”、“可水平扩展”、“可测试性高” 的应用系统 ——Newbe.Claptrap"}]},{"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":"本篇文章是该框架的一篇技术选文，属于技术构成的一部分。如果读者对该内容感兴趣，欢迎转发、评论、收藏文章以及项目。您的支持是促进项目成功的关键。"}]},{"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":"GitHub 项目地址："},{"type":"link","attrs":{"href":"https://github.com/newbe36524/Newbe.Claptrap","title":null},"content":[{"type":"text","text":"https://github.com/newbe36524/Newbe.Claptrap"}]}]},{"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":"Gitee 项目地址："},{"type":"link","attrs":{"href":"https://gitee.com/yks/Newbe.Claptrap","title":null},"content":[{"type":"text","text":"https://gitee.com/yks/Newbe.Claptrap"}]}]},{"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":"如果你对该项目感兴趣，你可以通过 github issues 提交您的看法。"}]},{"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":"如果您无法正常访问 github issue，您也可以发送邮件到 "},{"type":"link","attrs":{"href":"mailto:[email protected]","title":null},"content":[{"type":"text","text":"[email protected]"}]},{"type":"text","text":" 来参与我们的讨论。"}]},{"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":"点击链接 QQ 交流【Newbe.Claptrap】："},{"type":"link","attrs":{"href":"https://jq.qq.com/?_wv=1027&k=5uJGXf5","title":null},"content":[{"type":"text","text":"https://jq.qq.com/?_wv=1027&k=5uJGXf5"}]},{"type":"text","text":"。"}]},{"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":"​​​​​​​"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"本文作者： "},{"type":"text","text":"newbe36524"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"本文链接："},{"type":"text","text":" "},{"type":"link","attrs":{"href":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/","title":null},"content":[{"type":"text","text":"https://www.newbe.pro/Newbe.Claptrap/How-Many-RAMs-In-Used-While-There-Are-One-Hundred-Thousand-Users-Online/"}]}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"版权声明： "},{"type":"text","text":"本博客所有文章除特别声明外，均采用 "},{"type":"link","attrs":{"href":"https://creativecommons.org/licenses/by-nc-sa/4.0/","title":null},"content":[{"type":"text","text":"BY-NC-SA"}]},{"type":"text","text":" 许可协议。转载请注明出处！"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]}