猿论
首发于猿论
漫谈Java高并发方案

漫谈Java高并发方案

所有示例代码,请见/下载于
github.com/Wasabi1234/c





1 基本概念


1.1 并发

同时拥有两个或者多个线程,如果程序在单核处理器上运行多个线程将交替地换入或者换出内存,这些线程是同时“存在"的,每个线程都处于执行过程中的某个状态,如果运行在多核处理器上,此时,程序中的每个线程都将分配到一个处理器核上,因此可以同时运行.


1.2 高并发( High Concurrency)

互联网分布式系统架构设计中必须考虑的因素之一,通常是指,通过设计保证系统能够同时并行处理很多请求.


1.3 区别与联系

  • 并发: 多个线程操作相同的资源,保证线程安全,合理使用资源
  • 高并发:服务能同时处理很多请求,提高程序性能

2 CPU
2.1 CPU 多级缓存


  • 为什么需要CPU cache
    CPU的频率太快了,快到主存跟不上
    如此,在处理器时钟周期内,CPU常常需要等待主存,浪费资源。所以cache的出现,是为了缓解CPU和内存之间速度的不匹配问题(结构:cpu-> cache-> memory ).
  • CPU cache的意义
    1) 时间局部性
    如果某个数据被访问,那么在不久的将来它很可能被再次访问
    2) 空间局部性
    如果某个数据被访问,那么与它相邻的数据很快也可能被访问

2.2 缓存一致性(MESI)

用于保证多个 CPU cache 之间缓存共享数据的一致

  • M-modified被修改
    该缓存行只被缓存在该 CPU 的缓存中,并且是被修改过的,与主存中数据是不一致的,需在未来某个时间点写回主存,该时间是允许在其他CPU 读取主存中相应的内存之前,当这里的值被写入主存之后,该缓存行状态变为 E
  • E-exclusive独享
    缓存行只被缓存在该 CPU 的缓存中,未被修改过,与主存中数据一致
    可在任何时刻当被其他 CPU读取该内存时变成 S 态,被修改时变为 M态
  • S-shared共享
    该缓存行可被多个 CPU 缓存,与主存中数据一致
  • I-invalid无效


  • 乱序执行优化
    处理器为提高运算速度而做出违背代码原有顺序的优化
    并发的优势与风险



3 项目准备
3.1 项目初始化





3.2 并发模拟-Jmeter压测







3.3 并发模拟-代码CountDownLatch


Semaphore(信号量)


以上二者通常和线程池搭配

下面开始做并发模拟

package com.mmall.concurrency;

import com.mmall.concurrency.annoations.NotThreadSafe;
import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Semaphore;

/**
 * @author shishusheng
 * @date 18/4/1
 */
@Slf4j
@NotThreadSafe
public class ConcurrencyTest {

    /**
     * 请求总数
     */
    public static int clientTotal = 5000;

    /**
     * 同时并发执行的线程数
     */
    public static int threadTotal = 200;

    public static int count = 0;

    public static void main(String[] args) throws Exception {
        //定义线程池
        ExecutorService executorService = Executors.newCachedThreadPool();
        //定义信号量,给出允许并发的线程数目
        final Semaphore semaphore = new Semaphore(threadTotal);
        //统计计数结果
        final CountDownLatch countDownLatch = new CountDownLatch(clientTotal);
        //将请求放入线程池
        for (int i = 0; i < clientTotal ; i++) {
            executorService.execute(() -> {
                try {
                    //信号量的获取
                    semaphore.acquire();
                    add();
                    //释放
                    semaphore.release();
                } catch (Exception e) {
                    log.error("exception", e);
                }
                countDownLatch.countDown();
            });
        }
        countDownLatch.await();
        //关闭线程池
        executorService.shutdown();
        log.info("count:{}", count);
    }

    /**
     * 统计方法
     */
    private static void add() {
        count++;
    }
}

运行发现结果随机,所以非线程安全


4线程安全性


4.1 线程安全性

当多个线程访问某个类时,不管运行时环境采用何种调度方式或者这些进程将如何交替执行,并且在主调代码中不需要任何额外的同步或协同,这个类都能表现出正确的行为,那么就称这个类是线程安全的


4.2 原子性

4.2.1 Atomic 包

  • AtomicXXX:CAS,Unsafe.compareAndSwapInt
    提供了互斥访问,同一时刻只能有一个线程来对它进行操作
package com.mmall.concurrency.example.atomic;

import com.mmall.concurrency.annoations.ThreadSafe;
import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Semaphore;
import java.util.concurrent.atomic.AtomicLong;

/**
 * @author shishusheng
 */
@Slf4j
@ThreadSafe
public class AtomicExample2 {

    /**
     * 请求总数
     */
    public static int clientTotal = 5000;

    /**
     * 同时并发执行的线程数
     */
    public static int threadTotal = 200;

    /**
     * 工作内存
     */
    public static AtomicLong count = new AtomicLong(0);

    public static void main(String[] args) throws Exception {
        ExecutorService executorService = Executors.newCachedThreadPool();
        final Semaphore semaphore = new Semaphore(threadTotal);
        final CountDownLatch countDownLatch = new CountDownLatch(clientTotal);
        for (int i = 0; i < clientTotal ; i++) {
            executorService.execute(() -> {
                try {
                    System.out.println();
                    semaphore.acquire();
                    add();
                    semaphore.release();
                } catch (Exception e) {
                    log.error("exception", e);
                }
                countDownLatch.countDown();
            });
        }
        countDownLatch.await();
        executorService.shutdown();
        //主内存
        log.info("count:{}", count.get());
    }

    private static void add() {
        count.incrementAndGet();
        // count.getAndIncrement();
    }
}
package com.mmall.concurrency.example.atomic;

import com.mmall.concurrency.annoations.ThreadSafe;
import lombok.extern.slf4j.Slf4j;
import java.util.concurrent.atomic.AtomicReference;

/**
 * @author shishusheng
 * @date 18/4/3
 */
@Slf4j
@ThreadSafe
public class AtomicExample4 {

    private static AtomicReference count = new AtomicReference<>(0);

    public static void main(String[] args) {
        // 2
        count.compareAndSet(0, 2);
        // no
        count.compareAndSet(0, 1);
        // no
        count.compareAndSet(1, 3);
        // 4
        count.compareAndSet(2, 4);
        // no
        count.compareAndSet(3, 5); 
        log.info("count:{}", count.get());
    }
}



  • AtomicReference,AtomicReferenceFieldUpdater


  • AtomicBoolean


  • AtomicStampReference : CAS的 ABA 问题

4.2.2 锁
synchronized:依赖 JVM

  • 修饰代码块:大括号括起来的代码,作用于调用的对象
  • 修饰方法: 整个方法,作用于调用的对象


  • 修饰静态方法:整个静态方法,作用于所有对象


package com.mmall.concurrency.example.count;

import com.mmall.concurrency.annoations.ThreadSafe;
import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Semaphore;

/**
 * @author shishusheng
 */
@Slf4j
@ThreadSafe
public class CountExample3 {

    /**
     * 请求总数
     */
    public static int clientTotal = 5000;

    /**
     * 同时并发执行的线程数
     */
    public static int threadTotal = 200;

    public static int count = 0;

    public static void main(String[] args) throws Exception {
        ExecutorService executorService = Executors.newCachedThreadPool();
        final Semaphore semaphore = new Semaphore(threadTotal);
        final CountDownLatch countDownLatch = new CountDownLatch(clientTotal);
        for (int i = 0; i < clientTotal ; i++) {
            executorService.execute(() -> {
                try {
                    semaphore.acquire();
                    add();
                    semaphore.release();
                } catch (Exception e) {
                    log.error("exception", e);
                }
                countDownLatch.countDown();
            });
        }
        countDownLatch.await();
        executorService.shutdown();
        log.info("count:{}", count);
    }

    private synchronized static void add() {
        count++;
    }
}

synchronized 修正计数类方法

  • 修饰类:括号括起来的部分,作用于所有对象
    子类继承父类的被 synchronized 修饰方法时,是没有 synchronized 修饰的!!!

Lock: 依赖特殊的 CPU 指令,代码实现


4.2.3 对比

  • synchronized: 不可中断锁,适合竞争不激烈,可读性好
  • Lock: 可中断锁,多样化同步,竞争激烈时能维持常态
  • Atomic: 竞争激烈时能维持常态,比Lock性能好; 只能同步一
    个值

4.3 可见性

一个线程对主内存的修改可以及时的被其他线程观察到

4.3.1 导致共享变量在线程间不可见的原因

  • 线程交叉执行
  • 重排序结合线程交叉执行
  • 共享变量更新后的值没有在工作内存与主存间及时更新

4.3.2 可见性之synchronized
JMM关于synchronized的规定

  • 线程解锁前,必须把共享变量的最新值刷新到主内存
  • 线程加锁时,将清空工作内存中共享变量的值,从而使
    用共享变量时需要从主内存中重新读取最新的值(加锁与解锁是同一把锁)

4.3.3 可见性之volatile
通过加入内存屏障和禁止重排序优化来实现

  • 对volatile变量写操作时,会在写操作后加入一条store
    屏障指令,将本地内存中的共享变量值刷新到主内存
  • 对volatile变量读操作时,会在读操作前加入一条load
    屏障指令,从主内存中读取共享变量




  • volatile使用
volatile boolean inited = false;

//线程1:
context = loadContext();
inited= true;

// 线程2:
while( !inited ){
    sleep();
}
doSomethingWithConfig(context)


4.4 有序性

一个线程观察其他线程中的指令执行顺序,由于指令重排序的存在,该观察结果一般杂乱无序

JMM允许编译器和处理器对指令进行重排序,但是重排序过程不会影响到单线程程序的执行,却会影响到多线程并发执行的正确性

4.4.1 happens-before 规则


5发布对象






5.1 安全发布对象






package com.mmall.concurrency.example.singleton;

import com.mmall.concurrency.annoations.NotThreadSafe;

/**
 * 懒汉模式 -》 双重同步锁单例模式
 * 单例实例在第一次使用时进行创建
 * @author shishusheng
 */
@NotThreadSafe
public class SingletonExample4 {

    /**
     * 私有构造函数
     */
    private SingletonExample4() {

    }

    // 1、memory = allocate() 分配对象的内存空间
    // 2、ctorInstance() 初始化对象
    // 3、instance = memory 设置instance指向刚分配的内存

    // JVM和cpu优化,发生了指令重排

    // 1、memory = allocate() 分配对象的内存空间
    // 3、instance = memory 设置instance指向刚分配的内存
    // 2、ctorInstance() 初始化对象

    /**
     * 单例对象
     */
    private static SingletonExample4 instance = null;

    /**
     * 静态的工厂方法
     *
     * @return
     */
    public static SingletonExample4 getInstance() {
        // 双重检测机制 // B
        if (instance == null) {        
            // 同步锁
            synchronized (SingletonExample4.class) { 
                if (instance == null) {
                    // A - 3
                    instance = new SingletonExample4(); 
                }
            }
        }
        return instance;
    }
}





6 AQS


6.1 介绍



  • 使用Node实现FIFO队列,可以用于构建锁或者其他同步装置的基础框架
  • 利用了一个int类型表示状态
  • 使用方法是继承
  • 子类通过继承并通过实现它的方法管理其状态{acquire 和release} 的方法操纵状态
  • 可以同时实现排它锁和共享锁模式(独占、共享)

同步组件

CountDownLatch

package com.mmall.concurrency.example.aqs;


import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

/**

  • @author shishusheng
    */
    @Slf4j
    public class CountDownLatchExample1 {
    private final static int threadCount = 200;
    public static void main(String[] args) throws Exception {
ExecutorService exec = Executors.newCachedThreadPool();

final CountDownLatch countDownLatch = new CountDownLatch(threadCount);

for (int i = 0; i < threadCount; i++) {
    final int threadNum = i;
    exec.execute(() -> {
        try {
            test(threadNum);
        } catch (Exception e) {
            log.error("exception", e);
        } finally {
            countDownLatch.countDown();
        }
    });
}
countDownLatch.await();
log.info("finish");
exec.shutdown();

}
private static void test(int threadNum) throws Exception {
Thread.sleep(100);
log.info("{}", threadNum);
Thread.sleep(100);
}
}

package com.mmall.concurrency.example.aqs;

import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;

/**

  • 指定时间内处理任务
  • @author shishusheng
  • */
    @Slf4j
    public class CountDownLatchExample2 {
    private final static int threadCount = 200;
    public static void main(String[] args) throws Exception {
ExecutorService exec = Executors.newCachedThreadPool();

final CountDownLatch countDownLatch = new CountDownLatch(threadCount);

for (int i = 0; i < threadCount; i++) {
    final int threadNum = i;
    exec.execute(() -> {
        try {
            test(threadNum);
        } catch (Exception e) {
            log.error("exception", e);
        } finally {
            countDownLatch.countDown();
        }
    });
}
countDownLatch.await(10, TimeUnit.MILLISECONDS);
log.info("finish");
exec.shutdown();

}
private static void test(int threadNum) throws Exception {
Thread.sleep(100);
log.info("{}", threadNum);
}
}

##Semaphore用法
![](https://upload-images.jianshu.io/upload_images/4685968-e6cbcd4254c642c5.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
![](https://upload-images.jianshu.io/upload_images/4685968-dbefbf2c76ad5a2a.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
![](https://upload-images.jianshu.io/upload_images/4685968-41f5f5a5fd135804.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
##CycliBarrier

package com.mmall.concurrency.example.aqs;

import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CyclicBarrier;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

/**

  • @author shishusheng
    */
    @Slf4j
    public class CyclicBarrierExample1 {
    private static CyclicBarrier barrier = new CyclicBarrier(5);
    public static void main(String[] args) throws Exception {
ExecutorService executor = Executors.newCachedThreadPool();

for (int i = 0; i < 10; i++) {
    final int threadNum = i;
    Thread.sleep(1000);
    executor.execute(() -> {
        try {
            race(threadNum);
        } catch (Exception e) {
            log.error("exception", e);
        }
    });
}
executor.shutdown();

}
private static void race(int threadNum) throws Exception {
Thread.sleep(1000);
log.info("{} is ready", threadNum);
barrier.await();
log.info("{} continue", threadNum);
}
}

![](https://upload-images.jianshu.io/upload_images/4685968-4fb51fa4926fd70e.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)

package com.mmall.concurrency.example.aqs;


import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.CyclicBarrier;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;

/**

  • @author shishusheng
    */
    @Slf4j
    public class CyclicBarrierExample2 {
    private static CyclicBarrier barrier = new CyclicBarrier(5);
    public static void main(String[] args) throws Exception {
ExecutorService executor = Executors.newCachedThreadPool();

for (int i = 0; i < 10; i++) {
    final int threadNum = i;
    Thread.sleep(1000);
    executor.execute(() -> {
        try {
            race(threadNum);
        } catch (Exception e) {
            log.error("exception", e);
        }
    });
}
executor.shutdown();

}
private static void race(int threadNum) throws Exception {
Thread.sleep(1000);
log.info("{} is ready", threadNum);
try {
barrier.await(2000, TimeUnit.MILLISECONDS);
} catch (Exception e) {
log.warn("BarrierException", e);
}
log.info("{} continue", threadNum);
}
}

![await 超时导致程序抛异常](https://upload-images.jianshu.io/upload_images/4685968-0f899c23531f8ee8.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)

package com.mmall.concurrency.example.aqs;


import lombok.extern.slf4j.Slf4j;

import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Semaphore;
/**

  • @author shishusheng
    */
    @Slf4j
    public class SemaphoreExample3 {
    private final static int threadCount = 20;
    public static void main(String[] args) throws Exception {
ExecutorService exec = Executors.newCachedThreadPool();

final Semaphore semaphore = new Semaphore(3);

for (int i = 0; i < threadCount; i++) {
    final int threadNum = i;
    exec.execute(() -> {
        try {
            // 尝试获取一个许可
            if (semaphore.tryAcquire()) {
                test(threadNum);
                // 释放一个许可
                semaphore.release();
            }
        } catch (Exception e) {
            log.error("exception", e);
        }
    });


}
exec.shutdown();
}
private static void test(int threadNum) throws Exception {
log.info("{}", threadNum);
Thread.sleep(1000);
}

}

#9 线程池
##9.1 newCachedThreadPool
![](https://upload-images.jianshu.io/upload_images/4685968-1122da7a48223ba1.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
##9.2 newFixedThreadPool
![](https://upload-images.jianshu.io/upload_images/4685968-0ea942bf12e5210f.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
##9.3 newSingleThreadExecutor
看出是顺序执行的
![](https://upload-images.jianshu.io/upload_images/4685968-989d59429f589403.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
##9.4 newScheduledThreadPool
![](https://upload-images.jianshu.io/upload_images/4685968-f7536ec7a1cf6ecc.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
![](https://upload-images.jianshu.io/upload_images/4685968-c90e09d5bfe707e6.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
#10 死锁
![](https://upload-images.jianshu.io/upload_images/4685968-461f6a4251ae8ca4.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)
![](https://upload-images.jianshu.io/upload_images/4685968-46d58773e597195f.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)


作者:一生只为虞美人

链接:imooc.com/article/detai

来源:慕课网

本文原创发布于慕课网 ,转载请注明出处,谢谢合作


推荐阅读:

【官方】手记栏目认证作者招募,长期有效,随时报名!_慕课手记

有奖征文005期 |人生路上得一良师,是何感受?

Python里的那些坑

高并发和高性能系统中锁的影响

如何利用Webp和http缓存节省30%的网络流量

编辑于 2018-07-27

文章被以下专栏收录

    “只整码,不整事”。程序员界依据所用的不同语言划分不同帮派,会出现各种自黑互黑现象,这个专栏希望能进行正向的知识传播和分享,不相互拆台,不互怼,和谐友好进行交流。