Understand the use of protobuf serialization protocol

Protobuf It is a platform independent serialization protocol launched by Google , Compared with the traditional serialization ,Protobuf A smaller , More flexible , Can effectively improve transmission efficiency , Reduce the bandwidth occupied during data transmission . Simply speaking ,Protobuf Has the following characteristics ：

Language has nothing to do 、 Platform independent .Protobuf Support Java、C++、Python multilingual , Multiple platforms
Efficient . Than XML smaller （3 ~ 10 times ）、 faster （20 ~ 100 times ）、 It's simpler
Extensibility 、 Compatibility is good. . You can update the data structure , Without affecting and destroying the original old program

This article takes Animal Take entity class as an example , feel protobuf Characteristics ;

public class Animal implements Serializable {
    

    private Long id;
    private String name;
    private List<String> actions;
    
	/** Omit get、set Method **/
}

One 、 Use Protobuf

establish AnimalProto.proto, The data types defined are as follows ：

package com.keduw.protobuf;

message Animal{
    
    //id
    required int64 id = 1;
    //name
    optional string name = 2;
    //actions
    repeated string actions = 3;
}

Protobuf Defines its own grammar rules .package Define what we finally generate Java The package to which the code belongs ,required Indicates that this field is required ,optional Indicates that the field is optional , and repeated Indicates that the field is continuous , Corresponding to Java Middle represents a set , It is convenient to define the order of fields Protobuf Deserialization of fields . The specific corresponding data types can be referred to as follows ：

Write it well AnimalProto.proto, It can be provided through the official website protoc.exe Generate corresponding Java Code （ If you can't find it, you can download it at the end of the article ）.

Two 、 Serialization and deserialization

The code previously generated by the tool （AnimalProto） The methods of serialization and deserialization have been encapsulated for us , We just need to call the corresponding method .

introduce Protobuf Dependence

<dependency>
    <groupId>com.google.protobuf</groupId>
    <artifactId>protobuf-java</artifactId>
    <version>2.4.1</version>
</dependency>

serialize ：

/** *  The call object is constructed Builder, Complete attribute assignment and serialization  * @return */
public static byte[] protobufSerializer(){
    
    AnimalProto.Animal.Builder builder = AnimalProto.Animal.newBuilder();
    builder.setId(1L);
    builder.setName(" Piglets ");
    List<String> actions = new ArrayList<>();
    actions.add("eat");
    actions.add("run");
    builder.addAllActions(actions);
    return builder.build().toByteArray();
}

Deserialization ：

/** *  By calling parseFrom Then deserialization is completed  * @param bytes * @return * @throws InvalidProtocolBufferException */
public static Animal deserialize(byte[] bytes) throws Exception {
    
    AnimalProto.Animal pAnimal = AnimalProto.Animal.parseFrom(bytes);
    Animal animal = new Animal();
    animal.setId(pAnimal.getId());
    animal.setName(pAnimal.getName());
    animal.setActions(pAnimal.getActionsList());
    return animal;
}

test ：

public static void main(String[] args) throws Exception {
    
    byte[] bytes = serializer();
    Animal animal = deserialize(bytes);
    System.out.println(animal);
}

You can see that it can be serialized and deserialized normally .

3、 ... and 、 Performance comparison

in addition , I have also made a separate comparison here Java Native serialization 、jackson and Protobuf Serialization of the same data Animal Data length after operation , You can see Protobuf The length after serialization is obviously small , so Protobuf The performance of the is still excellent .

as for Protobuf Why can we achieve these improvements , On the one hand, data mapping is completed by defining field order , So you don't have to save the field name , On the other hand , This has something to do with the data structure he defines himself , Different serialization methods are used for different data types , Those who are interested in the implementation principle can understand by themselves .

The source code of this article and *.proto File conversion tools ：【 download 】