Java实现生产消费者模型

1. 使用BlockingQueue

Java内置了许多BlockingQueue的实现，例如ArrayBlockingQueue，LinkedBlockingQueue等。我们可以通过这些内置的阻塞队列来实现生产者消费者模型。

下面是producer的实现：

public class Producer implements Runnable {
    private String name;
    private BlockingQueue<String> queue;

    public Producer(String name, BlockingQueue<String> queue) {
        this.name = name;
        this.queue = queue;
    }

    @Override
    public void run() {
        for (int i = 0; i < 100; i++) {
            try {
                queue.put("message " + i);
                System.out.println(name + " produces message " + i);
                TimeUnit.SECONDS.sleep(1); // producer以每秒1条的速率生产100条消息
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
}

下面是consumer的实现：

public class Consumer implements Runnable {
    private String name;
    private BlockingQueue<String> queue;

    public Consumer(String name, BlockingQueue<String> queue) {
        this.name = name;
        this.queue = queue;
    }

    @Override
    public void run() {
        while (true) {
            try {
                String message = queue.take();
                System.out.println(name + " consumes " + message);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
}

在Main类中启动2个producer和1个consumer：

public class Main {
    public static void main(String[] args) {
        BlockingQueue<String> queue = new ArrayBlockingQueue<String>(5);
        new Thread(new Producer("producer1", queue)).start();
        new Thread(new Producer("producer2", queue)).start();
        new Thread(new Consumer("consumer1", queue)).start();
    }
}

2. 使用wait/notify实现

我们也可以通过wait/notify机制实现一个简单的BlockingQueue，从而实现生产者消费者模型。

下面是wait/notify机制实现的SimpleBlockingQueue：

public class SimpleBlockingQueue<T> {
    private LinkedList<T> elements;
    private int capacity;
    private final Object lock = new Object();

    public SimpleBlockingQueue(int size) {
        elements = new LinkedList<>();
        this.capacity = size;
    }

    public void put(T element) throws InterruptedException {
        synchronized (lock) {
            while (elements.size() == capacity) {
                lock.wait();
            }
            elements.add(element);
            lock.notifyAll();
        }
    }

    public T take() throws InterruptedException {
        synchronized (lock) {
            while (elements.isEmpty()) {
                lock.wait();
            }
            T first = elements.pollFirst();
            lock.notifyAll();
            return first;
        }
    }
}

Producer和Consumer的代码可以保持不变，只需要在构造时传入一个SimpleBlockingQueue实例。

3. 使用lock/condition实现

通过Lock锁，我们也能实现等待/通知模式。

下面是使用condition.await/condition.signal机制实现的SimpleBlockingQueue：

public class SimpleBlockingQueue<T> {
    private LinkedList<T> elements;
    private int capacity;
    private final Lock lock = new ReentrantLock();
    private final Condition condition = lock.newCondition();

    public SimpleBlockingQueue(int size) {
        elements = new LinkedList<>();
        this.capacity = size;
    }

    public void put(T element) throws InterruptedException {
        lock.lock();
        try {
            while (elements.size() == capacity) {
                condition.await();
            }
            elements.add(element);
            condition.signalAll();
        } finally {
            lock.unlock();
        }
    }

    public T take() throws InterruptedException {
        lock.lock();
        try {
            while (elements.isEmpty()) {
                condition.await();
            }
            T first = elements.pollFirst();
            condition.signalAll();
            return first;
        } finally {
            lock.unlock();
        }
    }
}

我们也可以通过两个单独的condition开关来实现细粒度的等待/通知模型

public class SimpleBlockingQueue<T> {
    private LinkedList<T> elements;
    private int capacity;
    private final Lock lock = new ReentrantLock();
    private final Condition notFull = lock.newCondition();
    private final Condition notEmpty = lock.newCondition();
    
    public SimpleBlockingQueue(int size) {
        elements = new LinkedList<>();
        this.capacity = size;
    }

    public void put(T element) throws InterruptedException {
        lock.lock();
        try {
            while (elements.size() == capacity) {
                notFull.await();
            }
            elements.add(element);
            notEmpty.signalAll();
        } finally {
            lock.unlock();
        }
    }

    public T take() throws InterruptedException {
        lock.lock();
        try {
            while (elements.isEmpty()) {
                notEmpty.await();
            }
            T first = elements.pollFirst();
            notFull.signalAll();
            return first;
        } finally {
            lock.unlock();
        }
    }
}