Kryo序列化和ProroStruff序列化性能比较

Posted 2022-12-08 杨龙飞的博客

JavaBean类

package SeriazleCompare;

/**
 * Created by yang on 16-12-13.
 */
public class Student 
    private String name;
    private int age;
    private int time;
    private String xupt;

    public Student()

    
    public Student(String name, int age, int time, String xupt) 
        this.name = name;
        this.age = age;
        this.time = time;
        this.xupt = xupt;
    

    @Override
    public String toString() 
        return "Student" +
                "name='" + name + '\\'' +
                ", age=" + age +
                ", time=" + time +
                ", xupt='" + xupt + '\\'' +
                '';
    

KryoSerialization序列化

package SeriazleCompare;

import com.esotericsoftware.kryo.Kryo;
import com.esotericsoftware.kryo.Registration;
import com.esotericsoftware.kryo.io.Input;
import com.esotericsoftware.kryo.io.Output;


/**
 * Created by yang on 16-11-22.
 */
public class KryoSerialization 
    private Kryo kryo;
    private Registration registration = null;
    private Class<?> t;
    public KryoSerialization()
        kryo = new Kryo();
        kryo.setReferences(true);
    
    public void register(Class<?> T)
        //注册类
        t = T;
        registration = kryo.register(t);
    
    public byte[] Serialize(Object object)
        Output output = null;
        output = new Output(1,4096);
        kryo.writeObject(output,object);
        byte[] bt = output.toBytes();
        output.flush();
        return bt;
    
    public <t> t Deserialize(byte[] bt)
        Input input = null;
        input = new Input(bt);
        t res = (t)kryo.readObject(input,registration.getType());
        input.close();
        return res;
    

ProtoStruff

package SeriazleCompare;


import com.dyuproject.protostuff.LinkedBuffer;
import com.dyuproject.protostuff.ProtostuffIOUtil;
import com.dyuproject.protostuff.Schema;
import com.dyuproject.protostuff.runtime.RuntimeSchema;
import org.objenesis.Objenesis;
import org.objenesis.ObjenesisStd;

import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;

public class ProstruffSerialization 
    private static Map<Class<?>,Schema<?>> cachedSchema = new ConcurrentHashMap<>();
    private static Objenesis objenesis = new ObjenesisStd(true);
    public ProstruffSerialization()
    @SuppressWarnings("unchecked")
    private static <T> Schema<T> getSchema(Class<T> cls)
        Schema<T> schema = (Schema<T>) cachedSchema.get(cls);
        if(schema == null)
            schema = RuntimeSchema.createFrom(cls);
            if(schema != null)
                cachedSchema.put(cls,schema);
            
        
        return schema;
    
    //序列化
    @SuppressWarnings("unchecked")
    public static <T> byte[] serializer(T obj)
        Class<T> cls = (Class<T>) obj.getClass();
        LinkedBuffer buffer = LinkedBuffer.allocate(LinkedBuffer.DEFAULT_BUFFER_SIZE);
        try
            Schema<T> schema = getSchema(cls);
            return ProtostuffIOUtil.toByteArray(obj, schema, buffer);
        catch (Exception e) 
            throw new IllegalStateException(e.getMessage(), e);
        finally
            buffer.clear();
        
    
    //反序列化
    public static <T> T deserializer(byte[] data, Class<T>cls)
        try
            T message = (T)objenesis.newInstance(cls);
            Schema<T> schema = getSchema(cls);
            ProtostuffIOUtil.mergeFrom(data, message, schema);
            return message;
        catch (Exception e) 
            throw new IllegalStateException(e.getMessage(), e);
        
    

测试类

package SeriazleCompare;



import java.util.ArrayList;
import java.util.List;

/**
 * Created by yang on 16-12-13.
 */
public class KryoProstruffTest 
    private static ProstruffSerialization protostruff = new ProstruffSerialization();
    private static KryoSerialization kryo = new KryoSerialization();
    public static int SIZE=10;
    public static void main(String[] args) 

        List<byte[]> kryoList = new ArrayList<>();
        List<byte[]> protoStruffList = new ArrayList<>();
        kryo.register(Student.class);
        long start = System.currentTimeMillis();
        byte[] data=null;
        for(int i = 0;i < SIZE; ++i)
            Student student = new Student("yang"+i,21,32,"There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.");
            data = kryo.Serialize(student);
            kryoList.add(data);
        
        long end = System.currentTimeMillis() - start;

        System.out.println("kryo序列化的时间是:"+end+" ,大小是:"+data.length);


        long Dstart = System.currentTimeMillis();
        Student temp=null;
        for(byte[] a:kryoList)
            temp = kryo.Deserialize(a);
        
        long Dend = System.currentTimeMillis()-Dstart;
        System.out.println("Kryo反序列化的时间是:"+Dend+" "+ temp.toString());


        long Pstart = System.currentTimeMillis();
        for(int i=0; i<SIZE;++i)
            Student student = new Student("long"+i,21,32,"There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.There is no such thing as a great talent without great will - power.");
            data =  ProstruffSerialization.serializer(student);
            protoStruffList.add(data);
        
        long Pend = System.currentTimeMillis()-Pstart;
        System.out.println("protoStruff序列化的时间是:"+Pend+ " ,大小是:"+ data.length);

        long PDstart = System.currentTimeMillis();
        for(byte[] a:protoStruffList)
            temp = protostruff.deserializer(a,Student.class);
        
        long PDend =System.currentTimeMillis()-PDstart;
        System.out.println("protostruff反序列化的时间是:"+PDend+" "+temp.toString());

    

SIZE=10:
测试结果:

kryo序列化的时间是:3 ,大小是:284
Kryo反序列化的时间是:22
protoStruff序列化的时间是:86 ,大小是:286
protostruff反序列化的时间是:10

SIZE = 100
测试结果:

kryo序列化的时间是:8 ,大小是:285
Kryo反序列化的时间是:27
protoStruff序列化的时间是:88 ,大小是:287
protostruff反序列化的时间是:12

SIZE = 1000
测试结果:

kryo序列化的时间是:37 ,大小是:286
Kryo反序列化的时间是:60
protoStruff序列化的时间是:145 ,大小是:288
protostruff反序列化的时间是:28

SIZE = 10000;
测试结果:

kryo序列化的时间是:165 ,大小是:287
Kryo反序列化的时间是:138
protoStruff序列化的时间是:185 ,大小是:289
protostruff反序列化的时间是:65

SIZE =100000;
测试结果:

kryo序列化的时间是:427 ,大小是:288
Kryo反序列化的时间是:332
protoStruff序列化的时间是:173 ,大小是:290
protostruff反序列化的时间是:154

SIZE = 1000000
测试结果:

kryo序列化的时间是:8383 ,大小是:289
Kryo反序列化的时间是:6668
protoStruff序列化的时间是:5783 ,大小是:291
protostruff反序列化的时间是:1303

结论:
序列化:
在序列化对象数量较少的情况下,Kryo序列化比protostruff 快.(SIZE 10000左右)　
在序列化对象数在100000个情况下,Kryo序列化速度迅速上升,protostruff比Kryo快.
反序列化:
不管在对象数量少与多的情况下,Kyro的反序列化的速度都低于protostruff.
大小:
同样属性大小的JavaBean序列化后,Kryo序列化后的size小于protostruff.