生产者消费者模式，基于阻塞队列

阅读更多

基于阻塞队列可以分容易实现生产者消费者模式

基本思路
生产者：负责生产对象，并放入阻塞队列
消费者：while true线程，阻塞的从阻塞队列中获取对象 并处理。


应用场景
服务器段分发器的处理、消息队列实现等等

核心组件
核心组件为JDK提供的阻塞队列，LinkedBlockingQueue


下面一个简单的例子
生产者

package com.gbcom.java.blockqueue;

import java.util.Random;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;

/**
 * 生产者线程。
 * 
 * @author SYZ
 * @date 2016-12-8 下午02:07:34
 * @version 1.0.0
 * @see com.gbcom.java.blockqueue.Producer
 */
public class Producer implements Runnable {

	private volatile boolean isRunning = true;
	private BlockingQueue queue;
	private static AtomicInteger count = new AtomicInteger();
	private static final int DEFAULT_RANGE_FOR_SLEEP = 1000;

	public Producer(BlockingQueue queue) {
		this.queue = queue;
	}

	public void run() {
		String data = null;
		Random r = new Random();

		System.out.println("启动生产者线程！");
		try {
			while (isRunning) {
				System.out.println("正在生产数据...");
				Thread.sleep(r.nextInt(DEFAULT_RANGE_FOR_SLEEP));

				data = "data:" + count.incrementAndGet();
				System.out.println("将数据：" + data + "放入队列...");
				if (queue.size() >= 5) {
					System.out
							.println("/***************** clear**********************/");
					queue.clear();
				}
				queue.put(data);
			}
		} catch (InterruptedException e) {
			e.printStackTrace();
			Thread.currentThread().interrupt();
		} finally {
			System.out.println("退出生产者线程！");
		}
	}

	public void stop() {
		isRunning = false;
	}

}

消费者

package com.gbcom.java.blockqueue;

import java.util.Random;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;

/**
 * 消费线程
 * 
 * @author SYZ
 * @date 2016-12-8 下午02:07:24
 * @version 1.0.0
 * @see com.gbcom.java.blockqueue.Consumer
 */
public class Consumer implements Runnable {
	private BlockingQueue queue;
	private static final int DEFAULT_RANGE_FOR_SLEEP = 1000;
	private static AtomicInteger count = new AtomicInteger();

	public Consumer(BlockingQueue queue) {
		this.queue = queue;
	}

	public void run() {
		System.out.println("启动消费者线程！");
		Random r = new Random();
		boolean isRunning = true;
		try {
			while (isRunning) {
				System.out.println("正从队列获取数据...");
				String data = queue.take();
				if (null != data) {
					System.out.println("拿到数据：" + data + "  : queue size = "
							+ queue.size());
					System.out.println(Thread.currentThread().getName()
							+ " - 正在消费数据：" + data + "::::consumer times="
							+ count.incrementAndGet());
					Thread.sleep(r.nextInt(DEFAULT_RANGE_FOR_SLEEP));
				} else {
					isRunning = false;
				}
			}
		} catch (InterruptedException e) {
			e.printStackTrace();
			Thread.currentThread().interrupt();
		} finally {
			System.out.println("退出消费者线程！");
		}
	}

}

客户端

package com.gbcom.java.blockqueue;

import java.util.concurrent.BlockingQueue;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.LinkedBlockingQueue;
import java.util.concurrent.ThreadPoolExecutor;

/**
 * 生产者消费者客户端
 * 
 * @author syz
 * @date 2014-7-2
 * @version v1.0.0
 * @see com.gbcom.java.blockqueue.BlockingQueueClient
 */
public class BlockingQueueClient {

	public static void main(String[] args) throws InterruptedException {
		// 声明一个容量为10的缓存队列
		BlockingQueue queue = new LinkedBlockingQueue(10);

		Producer producer1 = new Producer(queue);
		Producer producer2 = new Producer(queue);
		Producer producer3 = new Producer(queue);
		Consumer consumer = new Consumer(queue);
		// Consumer consumer2 = new Consumer(queue);
		// Consumer consumer3 = new Consumer(queue);

		// 借助Executors
		ThreadPoolExecutor service = (ThreadPoolExecutor) Executors
				.newCachedThreadPool();
		// 启动线程
		service.execute(producer1);
		service.execute(producer2);
		service.execute(producer3);
		// service.execute(consumer2);
		// service.execute(consumer3);
		service.execute(consumer);

		// 执行10s
		Thread.sleep(10 * 1000);

		System.out.println("active count = " + service.getActiveCount());
		// producer1.stop();
		// producer2.stop();
		// producer3.stop();
		Thread.sleep(2000);
		// 退出Executor
		service.shutdown();
	}
}