<h2>总览</h2>    <p>先来看一下 <strong>FlatBuffers</strong> 项目已经为我们提供了什么，而我们在将 <strong>FlatBuffers</strong> 用到我们的项目中时又需要做什么的整体流程。如下图：</p>    <p style="text-align:center"><img src="https://simg.open-open.com/show/09e8b7cca7c810c32b0c8a3606d25be4.jpg"></p>    <p>在使用 <strong>FlatBuffers</strong> 时，我们需要以特殊的格式定义我们的结构化数据，保存为 .fbs 文件。 <strong>FlatBuffers</strong> 项目为我们提供了编译器，可用于将 .fbs 文件编译为Java文件，C++文件等，以用于我们的项目。 <strong>FlatBuffers</strong> 编译器在我们的开发机，比如Ubuntu，Mac上运行。这些源代码文件是基于 <strong>FlatBuffers</strong> 提供的Java库生成的，同时我们也需要利用这个Java库的一些接口来序列化或解析数据。</p>    <p>我们将 <strong>FlatBuffers</strong> 编译器生成的Java文件及 <strong>FlatBuffers</strong> 的Java库导入我们的项目，就可以用 <strong>FlatBuffers</strong> 来对我们的结构化数据执行序列化和反序列化了。尽管每次手动执行 <strong>FlatBuffers</strong> 编译器生成Java文件非常麻烦，但不像 <strong>Protocol Buffers</strong> 那样，当前还没有Google官方提供的gradle插件可用。不过，我们这边开发了一个简单的 <strong>FlatBuffers</strong> gradle插件，后面会简单介绍一下，欢迎大家使用。</p>    <p>接下来我们更详细地看一下上面流程中的各个部分。</p>    <h2>下载、编译FlatBuffers编译器</h2>    <p>我们可以在如下位置：</p>    <pre>  <code class="language-java">https://github.com/google/flatbuffers/releases</code></pre>    <p>获取官方发布的打包好的版本。针对Windows平台有编译好的可执行安装文件，对其它平台还是打包的源文件。我们也可以指向clone repo的代码，进行手动编译。这里我们从GitHub上clone代码并手动编译编译器：</p>    <pre>  <code class="language-java">$ git clone https://github.com/google/flatbuffers.git  Cloning into 'flatbuffers'...  remote: Counting objects: 7340, done.  remote: Compressing objects: 100% (46/46), done.  remote: Total 7340 (delta 16), reused 0 (delta 0), pack-reused 7290  Receiving objects: 100% (7340/7340), 3.64 MiB | 115.00 KiB/s, done.  Resolving deltas: 100% (4692/4692), done.  Checking connectivity... done.</code></pre>    <p>下载代码之后，我们需要用cmake工具来为flatbuffers生成Makefile文件并编译：</p>    <pre>  <code class="language-java">$ cd flatbuffers/  $ cmake CMakeLists.txt   -- The C compiler identification is AppleClang 7.3.0.7030031  -- The CXX compiler identification is AppleClang 7.3.0.7030031  -- Check for working C compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc  -- Check for working C compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc -- works  -- Detecting C compiler ABI info  -- Detecting C compiler ABI info - done  -- Detecting C compile features  -- Detecting C compile features - done  -- Check for working CXX compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++  -- Check for working CXX compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++ -- works  -- Detecting CXX compiler ABI info  -- Detecting CXX compiler ABI info - done  -- Detecting CXX compile features  -- Detecting CXX compile features - done  -- Configuring done  -- Generating done  -- Build files have been written to: /Users/netease/Projects/OpenSource/flatbuffers  $ make && make install</code></pre>    <p>安装之后执行如下命令以确认已经装好：</p>    <pre>  <code class="language-java">$ flatc --version  flatc version 1.4.0 (Dec  7 2016)</code></pre>    <p>flatc没有为我们提供 <strong>--help</strong> 选项，不过加了错误的参数时这个工具会为我们展示详细的用法：</p>    <pre>  <code class="language-java">$ flatc --help  flatc: unknown commandline argument: --help  usage: flatc [OPTION]... FILE... [-- FILE...]    --binary     -b Generate wire format binaries for any data definitions.    --json       -t Generate text output for any data definitions.    --cpp        -c Generate C++ headers for tables/structs.    --go         -g Generate Go files for tables/structs.    --java       -j Generate Java classes for tables/structs.    --js         -s Generate JavaScript code for tables/structs.    --csharp     -n Generate C# classes for tables/structs.    --python     -p Generate Python files for tables/structs.    --php           Generate PHP files for tables/structs.    -o PATH            Prefix PATH to all generated files.    -I PATH            Search for includes in the specified path.    -M                 Print make rules for generated files.    --version          Print the version number of flatc and exit.    --strict-json      Strict JSON: field names must be / will be quoted,                       no trailing commas in tables/vectors.    --allow-non-utf8   Pass non-UTF-8 input through parser and emit nonstandard                       \x escapes in JSON. (Default is to raise parse error on                       non-UTF-8 input.)    --defaults-json    Output fields whose value is the default when                       writing JSON    --unknown-json     Allow fields in JSON that are not defined in the                       schema. These fields will be discared when generating                       binaries.    --no-prefix        Don't prefix enum values with the enum type in C++.    --scoped-enums     Use C++11 style scoped and strongly typed enums.                       also implies --no-prefix.    --gen-includes     (deprecated), this is the default behavior.                       If the original behavior is required (no include                       statements) use --no-includes.    --no-includes      Don't generate include statements for included                       schemas the generated file depends on (C++).    --gen-mutable      Generate accessors that can mutate buffers in-place.    --gen-onefile      Generate single output file for C#.    --gen-name-strings Generate type name functions for C++.    --escape-proto-ids Disable appending '_' in namespaces names.    --gen-object-api   Generate an additional object-based API.    --cpp-ptr-type T   Set object API pointer type (default std::unique_ptr)    --raw-binary       Allow binaries without file_indentifier to be read.                       This may crash flatc given a mismatched schema.    --proto            Input is a .proto, translate to .fbs.    --schema           Serialize schemas instead of JSON (use with -b)    --conform FILE     Specify a schema the following schemas should be                       an evolution of. Gives errors if not.    --conform-includes Include path for the schema given with --conform      PATH               FILEs may be schemas, or JSON files (conforming to preceding schema)  FILEs after the -- must be binary flatbuffer format files.  Output files are named using the base file name of the input,  and written to the current directory or the path given by -o.  example: flatc -c -b schema1.fbs schema2.fbs data.json</code></pre>    <h2>创建 .fbs 文件</h2>    <p>flatc支持将为 <strong>Protocol Buffers</strong> 编写的 .proto 文件转换为 .fbs 文件，如：</p>    <pre>  <code class="language-java">$ ls  addressbook.proto  $ flatc --proto addressbook.proto   $ ls -l  total 16  -rw-r--r--  1 netease  staff  431 12  7 17:21 addressbook.fbs  -rw-r--r--@ 1 netease  staff  486 12  1 15:18 addressbook.proto</code></pre>    <p>Protocol Buffers消息文件中的一些写法， <strong>FlatBuffers</strong> 编译器还不能很好的支持，如option java_package，option java_outer_classname，和嵌套类。这里我们基于 <strong>FlatBuffers</strong> 编译器转换的 .proto 文件来获得我们的 .fbs 文件：</p>    <pre>  <code class="language-java">// Generated from addressbook.proto    namespace com.example.tutorial;    enum PhoneType : int {    MOBILE = 0,    HOME = 1,    WORK = 2,  }    namespace com.example.tutorial;    table Person {    name:string (required);    id:int;    email:string;    phone:[com.example.tutorial._Person.PhoneNumber];  }    namespace com.example.tutorial._Person;    table PhoneNumber {    number:string (required);    type:int;  }    namespace com.example.tutorial;    table AddressBook {    person:[com.example.tutorial.Person];  }    root_type AddressBook;</code></pre>    <h2>编译 .fbs 文件</h2>    <p>可以通过如下命令编译 .fbs 文件：</p>    <pre>  <code class="language-java">$ flatc --java -o out addressbook.fbs</code></pre>    <p>--java用于指定编译的目标编程语言。-o 参数则用于指定输出文件的路径，如过没有提供则将当前目录用作输出目录。 <strong>FlatBuffers</strong> 编译器按照为不同的数据结构声明的namespace生成目录结构。对于上面的例子，会生成如下的这些文件：</p>    <pre>  <code class="language-java">$ find out  p.p1 {margin: 0.0px 0.0px 0.0px 0.0px; font: 11.0px Menlo}span.s1 {font-variant-ligatures: no-common-ligatures}    $ find out/  out/  out//com  out//com/example  out//com/example/tutorial  out//com/example/tutorial/_Person  out//com/example/tutorial/_Person/PhoneNumber.java  out//com/example/tutorial/AddressBook.java  out//com/example/tutorial/Person.java  out//com/example/tutorial/PhoneType.java</code></pre>    <h2>在Android项目中使用FlatBuffers</h2>    <p>我们将前面由 .fbs 文件生成的Java文件拷贝到我们的项目中。我们前面提到的， <strong>FlatBuffers</strong> 的Java库比较薄，当前并没有发不到jcenter这样的maven仓库中，因而我们需要将这部分代码也拷贝到我们的额项目中。 <strong>FlatBuffers</strong> 的Java库在其repo仓库的 java 目录下。引入这些文件之后，我们的代码结构如下：</p>    <p style="text-align:center"><img src="https://simg.open-open.com/show/e21068547bac05ca988f714e187613f4.jpg"></p>    <p>添加访问 <strong>FlatBuffers</strong> 的类：</p>    <pre>  <code class="language-java">package com.netease.volleydemo;    import com.example.tutorial.AddressBook;  import com.example.tutorial.Person;  import com.example.tutorial._Person.PhoneNumber;  import com.google.flatbuffers.FlatBufferBuilder;    import java.nio.ByteBuffer;    /**   * Created by hanpfei0306 on 16-12-5.   */    public class AddressBookFlatBuffers {      public static ByteBuffer encodeTest(String[] names) {          FlatBufferBuilder builder = new FlatBufferBuilder(0);            int[] personOffsets = new int[names.length];            for (int i = 0; i < names.length; ++ i) {              int name = builder.createString(names[i]);              int email = builder.createString("zhangsan@gmail.com");                int number1 = builder.createString("0157-23443276");              int type1 = 1;              int phoneNumber1 = PhoneNumber.createPhoneNumber(builder, number1, type1);                int number2 = builder.createString("136183667387");              int type2 = 0;              int phoneNumber2 = PhoneNumber.createPhoneNumber(builder, number2, type2);                int[] phoneNubers = new int[2];              phoneNubers[0] = phoneNumber1;              phoneNubers[1] = phoneNumber2;                int phoneNumbersPos = Person.createPhoneVector(builder, phoneNubers);                int person = Person.createPerson(builder, name, 13958235, email, phoneNumbersPos);                personOffsets[i] = person;          }          int persons = AddressBook.createPersonVector(builder, personOffsets);            AddressBook.startAddressBook(builder);          AddressBook.addPerson(builder, persons);          int eab = AddressBook.endAddressBook(builder);          builder.finish(eab);          ByteBuffer buf = builder.dataBuffer();            return buf;      }        public static ByteBuffer encodeTest(String[] names, int times) {          for (int i = 0; i < times - 1; ++ i) {              encodeTest(names);          }          return encodeTest(names);      }        public static AddressBook decodeTest(ByteBuffer byteBuffer) {          AddressBook addressBook = null;          addressBook = AddressBook.getRootAsAddressBook(byteBuffer);          return addressBook;      }        public static AddressBook decodeTest(ByteBuffer byteBuffer, int times) {          AddressBook addressBook = null;          for (int i = 0; i < times; ++ i) {              addressBook = decodeTest(byteBuffer);          }          return addressBook;      }  }</code></pre>    <h2>使用 flatbuf-gradle-plugin</h2>    <p>我们有开发一个 <strong>FlatBuffers</strong> 的gradle插件，以方便开发 。这个插件的设计有参考Google的protobuf-gradle-plugin，功能与用法也与protobuf-gradle-plugin类似。在这个项目中，我们也有为 <strong>FlatBuffers</strong> 的Java库创建一个module。</p>    <h2>编译并发布protobuf-gradle-plugin</h2>    <p>从github上下载代码：</p>    <pre>  <code class="language-java">$ git clone https://github.com/hanpfei/flatbuffers.git</code></pre>    <p>然后将代码导入Android Studio，将看到如下的代码结构：</p>    <p style="text-align:center"><img src="https://simg.open-open.com/show/5cade82e96a09d223f859206c3127948.jpg"></p>    <p>app 模块是一个demo程序，flatbuf-gradle-plugin 模块是 <strong>FlatBuffers</strong> 的gradle插件，而flatbuffers模块则是 <strong>FlatBuffers</strong> 的Java库。</p>    <p>为了使用 flatbuf-gradle-plugin，可以将插件发布到本地文件系统。这可以通过修改flatbuf-gradle-plugin/build.gradle来完成，修改 uploadArchives task 的 repository 指向本地文件系统，如：</p>    <pre>  <code class="language-java">uploadArchives {      repositories {          mavenDeployer {              pom.groupId = 'com.netease.hearttouch'              pom.artifactId = 'ht-flatbuf-gradle-plugin'              pom.version = '0.0.1-SNAPSHOT'              repository(url: 'file:///Users/netease/Projects/CorpProjects/ht-flatbuffers/app/plugin')          }      }  }</code></pre>    <p>执行uploadArchives task，编译并发布flatbuf-gradle-plugin到本地文件系统。</p>    <h2>应用flatbuf-gradle-plugin</h2>    <p>修改应用程序的 build.gradle 以应用 flatbuf-gradle-plugin 。</p>    <ol>     <li>为buildscript添加对 flatbuf-gradle-plugin 的依赖： <pre>  <code class="language-java">buildscript {   //目前先发布在本地，后面会通过maven进行引用   repositories {       maven {           url "file:///Users/netease/Projects/CorpProjects/ht-flatbuffers/app/plugin"       }       jcenter()       mavenCentral()   }   dependencies {       classpath 'com.netease.hearttouch:ht-flatbuf-gradle-plugin:0.0.1-SNAPSHOT'   }  }</code></pre> </li>     <li>在 apply plugin: 'com.android.application' 后面应用flatbuf的plugin： <pre>  <code class="language-java">apply plugin: 'com.android.application'  apply plugin: 'com.netease.flatbuf'</code></pre> </li>     <li> <p>添加flatbuf块，对flatbuf-gradle-plugin的执行做配置：</p> <pre>  <code class="language-java">flatbuf {   flatc {       path = '/usr/local/bin/flatc'   }     generateFlatTasks {       all().each { task ->           task.builtins {               remove java           }           task.builtins {               java { }           }       }   }  }</code></pre> <p>flatc 块用于配置 <strong>FlatBuffers</strong> 编译器，这里我们指定用我们之前手动编译的编译器。</p> <p>task.builtins 的块必不可少，这个块用于指定我们要为那些编程语言生成代码，这里我们为Java生成代码。</p> </li>     <li>指定 .fbs 文件的路径 <pre>  <code class="language-java">sourceSets {       main {           flat {               srcDir 'src/main/flat'           }       }   }</code></pre> 我们将 <strong>FlatBuffers</strong> 的IDL文件放在src/main/flat目录下。</li>    </ol>    <p>这样我们就不用再那么麻烦每次手动执行protoc了。</p>    <h2>FlatBuffers、Protobuf及JSON对比测试</h2>    <p>FlatBuffers相对于Protobuf的表现又如何呢？这里我们用数据说话，对比一下FlatBuffers格式、JSON格式与Protobuf的表现。测试同样用fastjson作为JSON的编码解码工具。</p>    <p>测试用的数据结构所有的数据结构，Protobuf相关的测试代码，及JSON的测试代码同 在Android中使用Protocol Buffers 一文所述，FlatBuffers的测试代码如下：</p>    <pre>  <code class="language-java">package hearttouch.netease.com.myapplication;    import com.example.tutorial.AddressBook;  import com.example.tutorial.Person;  import com.example.tutorial._Person.PhoneNumber;  import com.google.flatbuffers.FlatBufferBuilder;    import java.nio.ByteBuffer;    /**   * Created by hanpfei0306 on 16-12-5.   */    public class AddressBookFlatBuffers {      public static ByteBuffer encodeTest(String[] names) {          FlatBufferBuilder builder = new FlatBufferBuilder(0);            int[] personOffsets = new int[names.length];            for (int i = 0; i < names.length; ++ i) {              int name = builder.createString(names[i]);              int email = builder.createString("zhangsan@gmail.com");                int number1 = builder.createString("0157-23443276");              int type1 = 1;              int phoneNumber1 = PhoneNumber.createPhoneNumber(builder, number1, type1);                int number2 = builder.createString("136183667387");              int type2 = 0;              int phoneNumber2 = PhoneNumber.createPhoneNumber(builder, number2, type2);                int[] phoneNubers = new int[2];              phoneNubers[0] = phoneNumber1;              phoneNubers[1] = phoneNumber2;                int phoneNumbersPos = Person.createPhoneVector(builder, phoneNubers);                int person = Person.createPerson(builder, name, 13958235, email, phoneNumbersPos);                personOffsets[i] = person;          }          int persons = AddressBook.createPersonVector(builder, personOffsets);            AddressBook.startAddressBook(builder);          AddressBook.addPerson(builder, persons);          int eab = AddressBook.endAddressBook(builder);          builder.finish(eab);          ByteBuffer buf = builder.dataBuffer();            return buf;      }        public static ByteBuffer encodeTest(String[] names, int times) {          for (int i = 0; i < times - 1; ++ i) {              encodeTest(names);          }          return encodeTest(names);      }        public static AddressBook decodeTest(ByteBuffer byteBuffer) {          AddressBook addressBook = null;          addressBook = AddressBook.getRootAsAddressBook(byteBuffer);          return addressBook;      }        public static AddressBook decodeTest(ByteBuffer byteBuffer, int times) {          AddressBook addressBook = null;          for (int i = 0; i < times; ++ i) {              addressBook = decodeTest(byteBuffer);          }          return addressBook;      }  }</code></pre>    <p>通过如下的这段代码来执行测试：</p>    <pre>  <code class="language-java">private class ProtoTestTask extends AsyncTask<Void, Void, Void> {          private static final int BUFFER_LEN = 8192;            private void compress(InputStream is, OutputStream os)                  throws Exception {                GZIPOutputStream gos = new GZIPOutputStream(os);                int count;              byte data[] = new byte[BUFFER_LEN];              while ((count = is.read(data, 0, BUFFER_LEN)) != -1) {                  gos.write(data, 0, count);              }                gos.finish();              gos.close();          }            private int getCompressedDataLength(byte[] data) {              ByteArrayInputStream bais =new ByteArrayInputStream(data);              ByteArrayOutputStream baos = new ByteArrayOutputStream();                try {                  compress(bais, baos);              } catch (Exception e) {              }                return baos.toByteArray().length;          }            private void dumpDataLengthInfo(byte[] protobufData, String jsonData, ByteBuffer flatbufData) {              int compressedProtobufLength = getCompressedDataLength(protobufData);              int compressedJSONLength = getCompressedDataLength(jsonData.getBytes());              int compressedFlatbufLength = getCompressedDataLength(flatbufData.array());              Log.i(TAG, String.format("%-120s", "Data length"));              Log.i(TAG, String.format("%-20s%-20s%-20s%-20s%-20s%-20s", "Protobuf", "Protobuf (GZIP)",                      "JSON", "JSON (GZIP)", "Flatbuf", "Flatbuf (GZIP)"));              Log.i(TAG, String.format("%-20s%-20s%-20s%-20s%-20s%-20s",                      String.valueOf(protobufData.length), compressedProtobufLength,                      String.valueOf(jsonData.getBytes().length), compressedJSONLength,                      String.valueOf(flatbufData.array().length), compressedFlatbufLength));          }            private void doEncodeTest(String[] names, int times) {              long startTime = System.nanoTime();              byte[] protobufData = AddressBookProtobuf.encodeTest(names, times);              long protobufTime = System.nanoTime();              protobufTime = protobufTime - startTime;                startTime = System.nanoTime();              String jsonData = AddressBookJson.encodeTest(names, times);              long jsonTime = System.nanoTime();              jsonTime = jsonTime - startTime;                startTime = System.nanoTime();              ByteBuffer flatbufData = AddressBookFlatBuffers.encodeTest(names, times);              long flatbufTime = System.nanoTime();              flatbufTime = flatbufTime - startTime;                dumpDataLengthInfo(protobufData, jsonData, flatbufData);                Log.i(TAG, String.format("%-20s%-20s%-20s%-20s", "Encode Times", String.valueOf(times),                      "Names Length", String.valueOf(names.length)));                Log.i(TAG, String.format("%-20s%-20s%-20s%-20s%-20s%-20s",                      "ProtobufTime", String.valueOf(protobufTime),                      "JsonTime", String.valueOf(jsonTime),                      "FlatbufTime", String.valueOf(flatbufTime)));          }            private void doEncodeTest10(int times) {              doEncodeTest(TestUtils.sTestNames10, times);          }            private void doEncodeTest50(int times) {              doEncodeTest(TestUtils.sTestNames50, times);          }            private void doEncodeTest100(int times) {              doEncodeTest(TestUtils.sTestNames100, times);          }            private void doEncodeTest(int times) {              doEncodeTest10(times);              doEncodeTest50(times);              doEncodeTest100(times);          }            private void doDecodeTest(String[] names, int times) {              byte[] protobufBytes = AddressBookProtobuf.encodeTest(names);              ByteArrayInputStream bais = new ByteArrayInputStream(protobufBytes);              long startTime = System.nanoTime();              AddressBookProtobuf.decodeTest(bais, times);              long protobufTime = System.nanoTime();              protobufTime = protobufTime - startTime;                String jsonStr = AddressBookJson.encodeTest(names);              startTime = System.nanoTime();              AddressBookJson.decodeTest(jsonStr, times);              long jsonTime = System.nanoTime();              jsonTime = jsonTime - startTime;                ByteBuffer flatbufData = AddressBookFlatBuffers.encodeTest(names);              startTime = System.nanoTime();              AddressBookFlatBuffers.decodeTest(flatbufData, times);              long flatbufTime = System.nanoTime();              flatbufTime = flatbufTime - startTime;                Log.i(TAG, String.format("%-20s%-20s%-20s%-20s", "Decode Times", String.valueOf(times),                      "Names Length", String.valueOf(names.length)));              Log.i(TAG, String.format("%-20s%-20s%-20s%-20s%-20s%-20s",                      "ProtobufTime", String.valueOf(protobufTime),                      "JsonTime", String.valueOf(jsonTime),                      "FlatbufTime", String.valueOf(flatbufTime)));          }            private void doDecodeTest10(int times) {              doDecodeTest(TestUtils.sTestNames10, times);          }            private void doDecodeTest50(int times) {              doDecodeTest(TestUtils.sTestNames50, times);          }            private void doDecodeTest100(int times) {              doDecodeTest(TestUtils.sTestNames100, times);          }            private void doDecodeTest(int times) {              doDecodeTest10(times);              doDecodeTest50(times);              doDecodeTest100(times);          }            @Override          protected Void doInBackground(Void... params) {              TestUtils.initTest();              doEncodeTest(5000);                doDecodeTest(5000);              return null;          }            @Override          protected void onPostExecute(Void aVoid) {              super.onPostExecute(aVoid);          }      }</code></pre>    <p>这里我们执行3组编码测试及3组解码测试。对于编码测试，第一组的单个数据中包含10个Person，第二组的包含50个，第三组的包含100个，然后对每个数据分别执行5000次的编码操作。</p>    <p>对于解码测试，三组中单个数据同样包含10个Person、50个及100个，然后对每个数据分别执行5000次的解码码操作。</p>    <p>在Galaxy Nexus的Android 4.4.4 CM平台上执行上述测试，最终得到如下结果：</p>    <p>编码后数据长度对比 (Bytes)</p>    <table>     <thead>      <tr>       <th>Person个数</th>       <th>Protobuf</th>       <th>Protobuf（GZIP）</th>       <th>JSON</th>       <th>JSON（GZIP）</th>       <th>Flatbuf</th>       <th>Flatbuf（GZIP)</th>      </tr>     </thead>     <tbody>      <tr>       <td>10</td>       <td>860</td>       <td>290</td>       <td>1703</td>       <td>343</td>       <td>2048</td>       <td>521</td>      </tr>      <tr>       <td>50</td>       <td>4300</td>       <td>978</td>       <td>8463</td>       <td>1043</td>       <td>8192</td>       <td>1822</td>      </tr>      <tr>       <td>100</td>       <td>8600</td>       <td>1825</td>       <td>16913</td>       <td>1902</td>       <td>16384</td>       <td>3417</td>      </tr>     </tbody>    </table>    <p>相同的数据，经过编码，在压缩前FlatBuffers的数据长度和JSON的数据长度接近，而Protobuf的数据长度则只有前两者的大概一半。而在用GZIP压缩后，Protobuf的数据长度与JSON的接近，而FlatBuffers的数据长度则接近两者的两倍。</p>    <p>编码性能对比 (S)</p>    <table>     <thead>      <tr>       <th>Person个数</th>       <th>Protobuf</th>       <th>JSON</th>       <th>FlatBuffers</th>      </tr>     </thead>     <tbody>      <tr>       <td>10</td>       <td>6.000</td>       <td>8.952</td>       <td>12.464</td>      </tr>      <tr>       <td>50</td>       <td>26.847</td>       <td>45.782</td>       <td>56.752</td>      </tr>      <tr>       <td>100</td>       <td>50.602</td>       <td>73.688</td>       <td>108.426</td>      </tr>     </tbody>    </table>    <p>编码性能Protobuf相对于JSON有较大幅度的提高，而FlatBuffers则有较大幅度的降低。</p>    <p>解码性能对比 (S)</p>    <table>     <thead>      <tr>       <th>Person个数</th>       <th>Protobuf</th>       <th>JSON</th>       <th>FlatBuffers</th>      </tr>     </thead>     <tbody>      <tr>       <td>10</td>       <td>0.255</td>       <td>10.766</td>       <td>0.014</td>      </tr>      <tr>       <td>50</td>       <td>0.245</td>       <td>51.134</td>       <td>0.014</td>      </tr>      <tr>       <td>100</td>       <td>0.323</td>       <td>101.070</td>       <td>0.006</td>      </tr>     </tbody>    </table>    <p>解码性能方面，Protobuf相对于JSON，有着惊人的提升。Protobuf的解码时间几乎不随着数据长度的增长而有太大的增长，而JSON则随着数据长度的增加，解码所需要的时间也越来越长。而FlatBuffers则由于无需解码，在性能方面相对于前两者更有着非常大的提升。</p>    <p> </p>    <p> </p>    <p>来自：http://www.jianshu.com/p/3bac6bc80db7</p>    <p> </p>