如何在云数据库 Redis 版中进行管道传输

场景介绍


云数据库 Redis 版提供了与 Redis 相同的管道传输（pipeline）机制。管道（pipeline）将客户端 client 与服务器端的交互明确划分为单向的发送请求（Send Request）和接收响应（Receive Response）：用户可以将多个操作连续发给服务器，但在此期间服务器端并不对每个操作命令发送响应数据；全部请求发送完毕后用户关闭请求，开始接收响应获取每个操作命令的响应结果。
管道（pipeline）在某些场景下非常有用，比如有多个操作命令需要被迅速提交至服务器端，但用户并不依赖每个操作返回的响应结果，对结果响应也无需立即获得，那么管道就可以用来作为优化性能的批处理工具。性能提升的原因主要是减少了 TCP 连接中交互往返的开销。
不过在程序中使用管道请注意，使用 pipeline 时客户端将独占与服务器端的连接，此期间将不能进行其他“非管道”类型操作，直至 pipeline 被关闭；如果要同时执行其他操作，可以为 pipeline 操作单独建立一个连接，将其与常规操作分离开来。

代码示例1


性能对比 package pipeline.kvstore.aliyun.com;
import java.util.Date;
import redis.clients.jedis.Jedis;
import redis.clients.jedis.Pipeline;
public class RedisPipelinePerformanceTest {
        static final String host = "xxxxxx.m.cnhza.kvstore.aliyuncs.com";
        static final int port = 6379;
        static final String password = "password";
        public static void main(String[] args) {
            Jedis jedis = new Jedis(host, port);
                //ApsaraDB for Redis的实例密码
                String authString = jedis.auth(password);// password
                if (!authString.equals("OK")) {
                    System.err.println("AUTH Failed: " + authString);
                    jedis.close();
                    return;
                }
                //连续执行多次命令操作
                final int COUNT=5000;
                String key = "KVStore-Tanghan";
                // 1 ---不使用pipeline操作---
                jedis.del(key);//初始化key
                Date ts1 = new Date();
                for (int i = 0; i < COUNT; i++) {
                    //发送一个请求，并接收一个响应（Send Request and  Receive Response）
                    jedis.incr(key);
                }
                Date ts2 = new Date();
                System.out.println("不用Pipeline > value为:"+jedis.get(key)+" > 操作用时：" + (ts2.getTime() - ts1.getTime())+ "ms");
                //2 ----对比使用pipeline操作---
                jedis.del(key);//初始化key
                Pipeline p1 = jedis.pipelined();
                Date ts3 = new Date();
                for (int i = 0; i < COUNT; i++) {
                    //发出请求 Send Request
                    p1.incr(key);
                }
                //接收响应 Receive Response
                p1.sync();
                Date ts4 = new Date();
                System.out.println("使用Pipeline > value为:"+jedis.get(key)+" > 操作用时：" + (ts4.getTime() - ts3.getTime())+ "ms");
                jedis.close();
        }
    }


运行结果1


在输入了正确的云数据库 Redis 版实例访问地址和密码之后，运行以上 Java 程序，输出结果如下。从中可以看出使用 pipeline 的性能要快的多。 不用Pipeline > value为:5000 > 操作用时：5844ms
使用Pipeline > value为:5000 > 操作用时：78ms


代码示例2


在 Jedis 中使用管道（pipeline）时，对于响应数据（response）的处理有两种方式，请参考以下代码示例。 package pipeline.kvstore.aliyun.com;
import java.util.List;
import redis.clients.jedis.Jedis;
import redis.clients.jedis.Pipeline;
import redis.clients.jedis.Response;
    public class PipelineClientTest {
        static final String host = "xxxxxxxx.m.cnhza.kvstore.aliyuncs.com";
        static final int port = 6379;
        static final String password = "password";
        public static void main(String[] args) {
            Jedis jedis = new Jedis(host, port);
                // ApsaraDB for Redis的实例密码
                String authString = jedis.auth(password);// password
                if (!authString.equals("OK")) {
                    System.err.println("AUTH Failed: " + authString);
                    jedis.close();
                    return;
                }
                String key = "KVStore-Test1";
                jedis.del(key);//初始化
                // -------- 方法1
                Pipeline p1 = jedis.pipelined();
                System.out.println("-----方法1-----");
                for (int i = 0; i < 5; i++) {
                    p1.incr(key);
                    System.out.println("Pipeline发送请求");
                }
                // 发送请求完成，开始接收响应
                System.out.println("发送请求完成，开始接收响应");
                List<Object> responses = p1.syncAndReturnAll();
                if (responses == null || responses.isEmpty()) {
                    jedis.close();
                    throw new RuntimeException("Pipeline error: 没有接收到响应");
                }
                for (Object resp : responses) {
                    System.out.println("Pipeline接收响应Response: " + resp.toString());
                }
                System.out.println();
                //-------- 方法2
                System.out.println("-----方法2-----");
                jedis.del(key);//初始化
                Pipeline p2 = jedis.pipelined();  
                //需要先声明Response
                Response<Long> r1 = p2.incr(key);
                System.out.println("Pipeline发送请求");
                Response<Long> r2 = p2.incr(key);
                System.out.println("Pipeline发送请求");
                Response<Long> r3 = p2.incr(key);
                System.out.println("Pipeline发送请求");
                Response<Long> r4 = p2.incr(key);  
                System.out.println("Pipeline发送请求");
                Response<Long> r5 = p2.incr(key);
                System.out.println("Pipeline发送请求");
                try{  
                    r1.get();  //此时还未开始接收响应，所以此操作会出错
                }catch(Exception e){  
                    System.out.println(" <<< Pipeline error：还未开始接收响应  >>> ");  
                }  
             // 发送请求完成，开始接收响应
                System.out.println("发送请求完成，开始接收响应");
                p2.sync();  
                System.out.println("Pipeline接收响应Response: " + r1.get());  
                System.out.println("Pipeline接收响应Response: " + r2.get());  
                System.out.println("Pipeline接收响应Response: " + r3.get());
                System.out.println("Pipeline接收响应Response: " + r4.get());
                System.out.println("Pipeline接收响应Response: " + r5.get());
                jedis.close();
            }
    }


运行结果2


在输入了正确的云数据库 Redis 版实例访问地址和密码之后，运行以上 Java 程序，输出结果如下: -----方法1-----
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
发送请求完成，开始接收响应
Pipeline接收响应Response: 1
Pipeline接收响应Response: 2
Pipeline接收响应Response: 3
Pipeline接收响应Response: 4
Pipeline接收响应Response: 5
-----方法2-----
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
Pipeline发送请求
<<< Pipeline error：还未开始接收响应  >>>
发送请求完成，开始接收响应
Pipeline接收响应Response: 1
Pipeline接收响应Response: 2
Pipeline接收响应Response: 3
Pipeline接收响应Response: 4
Pipeline接收响应Response: 5