關於Spring Data redis幾種對象序列化的比較

 
    redis雖然提供了對list set hash等數據類型的支持，但是沒有提供對POJO對象的支持，底層都是把對象序列化后再以字符串的方式存儲的。因此，Spring data提供了若干個Serializer，主要包括： 

• JacksonJsonRedisSerializer
• JdkSerializationRedisSerializer
• OxmSerializer

   參見：http://static.springsource.org/spring-data/data-keyvalue/docs/1.0.x/api/ 

   這里，我第一是想測試一下三者的使用，第二是想看看它們的使用效果。 

准備工作 

下載源碼   
我直接在《Spring Data》書的源碼基礎上改，從這下載書的源碼：https://github.com/SpringSource/spring-data-book 

打開redis子項目，由於是以Maven組織的，所以不用關心包的問題。 

添加一個測試的Entity 

由於我們希望測試使用Redis保存POJO對象，因此我們在com.oreilly.springdata.redis包下創建一個User對象，如下所示： 

package com.oreilly.springdata.redis;

import javax.xml.bind.annotation.XmlAccessType;
import javax.xml.bind.annotation.XmlAccessorType;
import javax.xml.bind.annotation.XmlAttribute;
import javax.xml.bind.annotation.XmlRootElement;
import java.io.Serializable;

/**
 * @author : stamen
 * @date: 13-7-16
 */
@XmlAccessorType(XmlAccessType.FIELD)
@XmlRootElement(name = "user")
public class User implements Serializable {

    @XmlAttribute
    private String userName;

    @XmlAttribute
    private int age;

    public String getUserName() {
        return userName;
    }

    public void setUserName(String userName) {
        this.userName = userName;
    }

    public int getAge() {
        return age;
    }

    public void setAge(int age) {
        this.age = age;
    }
}

  
   由於后面，我們需要使用OXM及Jackson將進行對象序列，為了控制對象的序列化，因此打上了JSR 175注解。 

更改ApplicationConfig 

   ApplicationConfig是Spring容器的配置類，要根據你的環境進行更改，我的更改為： 

package com.oreilly.springdata.redis;

import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.data.redis.connection.RedisConnectionFactory;
import org.springframework.data.redis.connection.jedis.JedisConnectionFactory;
import org.springframework.data.redis.core.RedisTemplate;
import org.springframework.data.redis.serializer.OxmSerializer;
import org.springframework.data.redis.serializer.RedisSerializer;
import org.springframework.data.redis.serializer.SerializationException;
import org.springframework.oxm.jaxb.Jaxb2Marshaller;

import javax.xml.bind.JAXBContext;
import javax.xml.bind.Marshaller;
import java.util.HashMap;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;

/**
 * @author Jon Brisbin
 */
@Configuration
public abstract class ApplicationConfig {

    @Bean
    public RedisConnectionFactory redisConnectionFactory() {
        JedisConnectionFactory cf = new JedisConnectionFactory();
        cf.setHostName("10.188.182.140");
        cf.setPort(6379);
        cf.setPassword("superman");
        cf.afterPropertiesSet();
        return cf;
    }

    @Bean
    public RedisTemplate redisTemplate() {
        RedisTemplate rt = new RedisTemplate();
        rt.setConnectionFactory(redisConnectionFactory());
        return rt;
    }

    private static Map<Class, JAXBContext> jaxbContextHashMap = new ConcurrentHashMap<Class, JAXBContext>();

    @Bean
    public OxmSerializer oxmSerializer() throws Throwable{
        Jaxb2Marshaller jaxb2Marshaller = new Jaxb2Marshaller();
        Map<String,Object> properties = new HashMap<String, Object>();//創建映射，用於設置Marshaller屬性
        properties.put(Marshaller.JAXB_FORMATTED_OUTPUT, Boolean.TRUE);   //放置xml自動縮進屬性
        properties.put(Marshaller.JAXB_ENCODING,"utf-8");   //放置xml自動縮進屬性
        jaxb2Marshaller.setClassesToBeBound(User.class);//映射的xml類放入JAXB環境中
        jaxb2Marshaller.setMarshallerProperties(properties);//設置Marshaller屬性
        return  new OxmSerializer(jaxb2Marshaller,jaxb2Marshaller);
    }

    public static enum StringSerializer implements RedisSerializer<String> {
        INSTANCE;

        @Override
        public byte[] serialize(String s) throws SerializationException {
            return (null != s ? s.getBytes() : new byte[0]);
        }

        @Override
        public String deserialize(byte[] bytes) throws SerializationException {
            if (bytes.length > 0) {
                return new String(bytes);
            } else {
                return null;
            }
        }
    }

    public static enum LongSerializer implements RedisSerializer<Long> {
        INSTANCE;

        @Override
        public byte[] serialize(Long aLong) throws SerializationException {
            if (null != aLong) {
                return aLong.toString().getBytes();
            } else {
                return new byte[0];
            }
        }

        @Override
        public Long deserialize(byte[] bytes) throws SerializationException {
            if (bytes.length > 0) {
                return Long.parseLong(new String(bytes));
            } else {
                return null;
            }
        }
    }


    public static enum IntSerializer implements RedisSerializer<Integer> {
        INSTANCE;

        @Override
        public byte[] serialize(Integer i) throws SerializationException {
            if (null != i) {
                return i.toString().getBytes();
            } else {
                return new byte[0];
            }
        }

        @Override
        public Integer deserialize(byte[] bytes) throws SerializationException {
            if (bytes.length > 0) {
                return Integer.parseInt(new String(bytes));
            } else {
                return null;
            }
        }
    }

}

   1)redisConnectionFactory()配置了如何連接Redsi服務器（如何安裝Redis,參見：http://redis.io/download） 
   2）oxmSerializer()是我新增的，用於定義一個基於Jaxb2Marshaller的OxmSerializer Bean(后面將會用到) 


編寫測試用例 

    打開KeyValueSerializersTest，我們幾個額外的測試用例都將寫在該測試類中： 

使用JdkSerializationRedisSerializer序列化 

@Test
public void testJdkSerialiable() {
    RedisTemplate<String, Serializable> redis = new RedisTemplate<String, Serializable>();
    redis.setConnectionFactory(connectionFactory);
    redis.setKeySerializer(ApplicationConfig.StringSerializer.INSTANCE);
    redis.setValueSerializer(new JdkSerializationRedisSerializer());
    redis.afterPropertiesSet();

    ValueOperations<String, Serializable> ops = redis.opsForValue();

    User user1 = new User();
    user1.setUserName("user1");
    user1.setAge(20);

    String key1 = "users/user1";
    User user11 = null;

    long begin = System.currentTimeMillis();
    for (int i = 0; i < 100; i++) {
        ops.set(key1,user1);
        user11 = (User)ops.get(key1);
    }
    long time = System.currentTimeMillis() - begin;
    System.out.println("jdk time:"+time);
    assertThat(user11.getUserName(),is("user1"));
}

   JdkSerializationRedisSerializer支持對所有實現了Serializable的類進行序列化。運行該測試用例，我們通過redis-cli 通過“users/user1”鍵可以查看到對應的值，內容如下： 

引用

redis 127.0.0.1:6379> get users/user1 
"\xac\xed\x00\x05sr\x00!com.oreilly.springdata.redis.User\xb1\x1c \n\xcd\xed%\xd8\x02\x00\x02I\x00\x03ageL\x00\buserNamet\x00\x12Ljava/lang/String;xp\x00\x00\x00\x14t\x00\x05user1"

通過strlen查看對應的字符長度： 

引用

redis 127.0.0.1:6379> strlen users/user1 
(integer) 104 

上面的代碼共進行了100次的存儲和獲取，其所花時間如下（毫秒）： 

引用

jdk time:266

使用JacksonJsonRedisSerializer序列化 

@Test
public void testJacksonSerialiable() {
    RedisTemplate<String, Object> redis = new RedisTemplate<String, Object>();
    redis.setConnectionFactory(connectionFactory);
    redis.setKeySerializer(ApplicationConfig.StringSerializer.INSTANCE);
    redis.setValueSerializer(new JacksonJsonRedisSerializer<User>(User.class));
    redis.afterPropertiesSet();

    ValueOperations<String, Object> ops = redis.opsForValue();

    User user1 = new User();
    user1.setUserName("user1");
    user1.setAge(20);

    User user11 = null;
    String key1 = "json/user1";

    long begin = System.currentTimeMillis();
    for (int i = 0; i < 100; i++) {
        ops.set(key1,user1);
        user11 = (User)ops.get(key1);
    }
    long time = System.currentTimeMillis() - begin;

    System.out.println("json time:"+time);
    assertThat(user11.getUserName(),is("user1"));
}

    運行后，查看redis的內容及內容長度： 

引用

redis 127.0.0.1:6379> get json/user1 
"{\"userName\":\"user1\",\"age\":20}" 
redis 127.0.0.1:6379> strlen json/user1 
(integer) 29 

    執行花費時間為： 

引用

    json time:224 

使用OxmSerialiable序列化 

@Test
public void testOxmSerialiable() throws Throwable{
    RedisTemplate<String, Object> redis = new RedisTemplate<String, Object>();
    redis.setConnectionFactory(connectionFactory);
    redis.setKeySerializer(ApplicationConfig.StringSerializer.INSTANCE);

    redis.setValueSerializer(oxmSerializer);
    redis.afterPropertiesSet();

    ValueOperations<String, Object> ops = redis.opsForValue();

    User user1 = new User();
    user1.setUserName("user1");
    user1.setAge(20);


    User user11 = null;
    String key1 = "oxm/user1";

    long begin = System.currentTimeMillis();
    for (int i = 0; i < 100; i++) {
        ops.set(key1,user1);
        user11 = (User)ops.get(key1);
    }
    long time = System.currentTimeMillis() - begin;

    System.out.println("oxm time:"+time);
    assertThat(user11.getUserName(),is("user1"));
}

    運行后，查看redis的內容及內容長度： 

引用

redis 127.0.0.1:6379> get oxm/user1 
"<?xml version=\"1.0\" encoding=\"utf-8\" standalone=\"yes\"?>\n<user age=\"20\" userName=\"user1\"/>\n" 
redis 127.0.0.1:6379> strlen oxm/user1 
(integer) 90 

    執行花費時間為： 

引用

    oxm time:335 

小結 

    從執行時間上來看，JdkSerializationRedisSerializer是最高效的（畢竟是JDK原生的），但是是序列化的結果字符串是最長的。JSON由於其數據格式的緊湊性，序列化的長度是最小的，時間比前者要多一些。而OxmSerialiabler在時間上看是最長的（當時和使用具體的Marshaller有關）。所以個人的選擇是傾向使用JacksonJsonRedisSerializer作為POJO的序列器。