`
walksing
  • 浏览: 217120 次
  • 性别: Icon_minigender_1
  • 来自: 深圳
社区版块
存档分类
最新评论
阅读更多


我们谈一下实际的场景吧。我们在开发中,有如下场景

a) 关闭空闲连接。服务器中,有很多客户端的连接,空闲一段时间之后需要关闭之。
b) 缓存。缓存中的对象,超过了空闲时间,需要从缓存中移出。
c) 任务超时处理。在网络协议滑动窗口请求应答式交互时,处理超时未响应的请求。

一种笨笨的办法就是,使用一个后台线程,遍历所有对象,挨个检查。这种笨笨的办法简单好用,但是对象数量过多时,可能存在性能问题,检查间隔时间不好设置,间隔时间过大,影响精确度,多小则存在效率问题。而且做不到按超时的时间顺序处理。

这场景,使用DelayQueue最适合了。

DelayQueue是java.util.concurrent中提供的一个很有意思的类。很巧妙,非常棒!但是java doc和Java SE 5.0的source中都没有提供Sample。我最初在阅读ScheduledThreadPoolExecutor源码时,发现DelayQueue的妙用。随后在实际工作中,应用在session超时管理,网络应答通讯协议的请求超时处理。

本文将会对DelayQueue做一个介绍,然后列举应用场景。并且提供一个Delayed接口的实现和Sample代码。

DelayQueue是一个BlockingQueue,其特化的参数是Delayed。(不了解BlockingQueue的同学,先去了解BlockingQueue再看本文)
Delayed扩展了Comparable接口,比较的基准为延时的时间值,Delayed接口的实现类getDelay的返回值应为固定值(final)。DelayQueue内部是使用PriorityQueue实现的。

DelayQueue = BlockingQueue + PriorityQueue + Delayed

DelayQueue的关键元素BlockingQueue、PriorityQueue、Delayed。可以这么说,DelayQueue是一个使用优先队列(PriorityQueue)实现的BlockingQueue,优先队列的比较基准值是时间。

他们的基本定义如下
public interface Comparable<T> {
    public int compareTo(T o);
}
public interface Delayed extends Comparable<Delayed> {
    long getDelay(TimeUnit unit);
}
public class DelayQueue<E extends Delayed> implements BlockingQueue<E> {
    private final PriorityQueue<E> q = new PriorityQueue<E>();
}
DelayQueue内部的实现使用了一个优先队列。当调用DelayQueue的offer方法时,把Delayed对象加入到优先队列q中。如下:
public boolean offer(E e) {
    final ReentrantLock lock = this.lock;
    lock.lock();
    try {
        E first = q.peek();
        q.offer(e);
        if (first == null || e.compareTo(first) < 0)
            available.signalAll();
        return true;
    } finally {
        lock.unlock();
    }
}
DelayQueue的take方法,把优先队列q的first拿出来(peek),如果没有达到延时阀值,则进行await处理。如下:
public E take() throws InterruptedException {
    final ReentrantLock lock = this.lock;
    lock.lockInterruptibly();
    try {
        for (;;) {
            E first = q.peek();
            if (first == null) {
                available.await();
            } else {
                long delay =  first.getDelay(TimeUnit.NANOSECONDS);
                if (delay > 0) {
                    long tl = available.awaitNanos(delay);
                } else {
                    E x = q.poll();
                    assert x != null;
                    if (q.size() != 0)
                        available.signalAll(); // wake up other takers
                    return x;

                }
            }
        }
    } finally {
        lock.unlock();
    }
}
-------------------

以下是Sample,是一个缓存的简单实现。共包括三个类Pair、DelayItem、Cache。如下:

public class Pair<K, V> {
    public K first;

    public V second;
   
    public Pair() {}
   
    public Pair(K first, V second) {
        this.first = first;
        this.second = second;
    }
}
--------------
以下是Delayed的实现
import java.util.concurrent.Delayed;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicLong;

public class DelayItem<T> implements Delayed {
    /** Base of nanosecond timings, to avoid wrapping */
    private static final long NANO_ORIGIN = System.nanoTime();

    /**
     * Returns nanosecond time offset by origin
     */
    final static long now() {
        return System.nanoTime() - NANO_ORIGIN;
    }

    /**
     * Sequence number to break scheduling ties, and in turn to guarantee FIFO order among tied
     * entries.
     */
    private static final AtomicLong sequencer = new AtomicLong(0);

    /** Sequence number to break ties FIFO */
    private final long sequenceNumber;

    /** The time the task is enabled to execute in nanoTime units */
    private final long time;

    private final T item;

    public DelayItem(T submit, long timeout) {
        this.time = now() + timeout;
        this.item = submit;
        this.sequenceNumber = sequencer.getAndIncrement();
    }

    public T getItem() {
        return this.item;
    }

    public long getDelay(TimeUnit unit) {
        long d = unit.convert(time - now(), TimeUnit.NANOSECONDS);
        return d;
    }

    public int compareTo(Delayed other) {
        if (other == this) // compare zero ONLY if same object
            return 0;
        if (other instanceof DelayItem) {
            DelayItem x = (DelayItem) other;
            long diff = time - x.time;
            if (diff < 0)
                return -1;
            else if (diff > 0)
                return 1;
            else if (sequenceNumber < x.sequenceNumber)
                return -1;
            else
                return 1;
        }
        long d = (getDelay(TimeUnit.NANOSECONDS) - other.getDelay(TimeUnit.NANOSECONDS));
        return (d == 0) ? 0 : ((d < 0) ? -1 : 1);
    }
}



以下是Cache的实现,包括了put和get方法,还包括了可执行的main函数。
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.DelayQueue;
import java.util.concurrent.TimeUnit;
import java.util.logging.Level;
import java.util.logging.Logger;

public class Cache<K, V> {
    private static final Logger LOG = Logger.getLogger(Cache.class.getName());

    private ConcurrentMap<K, V> cacheObjMap = new ConcurrentHashMap<K, V>();

    private DelayQueue<DelayItem<Pair<K, V>>> q = new DelayQueue<DelayItem<Pair<K, V>>>();

    private Thread daemonThread;

    public Cache() {

        Runnable daemonTask = new Runnable() {
            public void run() {
                daemonCheck();
            }
        };

        daemonThread = new Thread(daemonTask);
        daemonThread.setDaemon(true);
        daemonThread.setName("Cache Daemon");
        daemonThread.start();
    }

    private void daemonCheck() {

        if (LOG.isLoggable(Level.INFO))
            LOG.info("cache service started.");

        for (;;) {
            try {
                DelayItem<Pair<K, V>> delayItem = q.take();
                if (delayItem != null) {
                    // 超时对象处理
                    Pair<K, V> pair = delayItem.getItem();
                    cacheObjMap.remove(pair.first, pair.second); // compare and remove
                }
            } catch (InterruptedException e) {
                if (LOG.isLoggable(Level.SEVERE))
                    LOG.log(Level.SEVERE, e.getMessage(), e);
                break;
            }
        }

        if (LOG.isLoggable(Level.INFO))
            LOG.info("cache service stopped.");
    }

    // 添加缓存对象
    public void put(K key, V value, long time, TimeUnit unit) {
        V oldValue = cacheObjMap.put(key, value);
        if (oldValue != null)
            q.remove(key);

        long nanoTime = TimeUnit.NANOSECONDS.convert(time, unit);
        q.put(new DelayItem<Pair<K, V>>(new Pair<K, V>(key, value), nanoTime));
    }

    public V get(K key) {
        return cacheObjMap.get(key);
    }

    // 测试入口函数
    public static void main(String[] args) throws Exception {
        Cache<Integer, String> cache = new Cache<Integer, String>();
        cache.put(1, "aaaa", 3, TimeUnit.SECONDS);

        Thread.sleep(1000 * 2);
        {
            String str = cache.get(1);
            System.out.println(str);
        }

        Thread.sleep(1000 * 2);
        {
            String str = cache.get(1);
            System.out.println(str);
        }
    }
}
运行Sample,main函数执行的结果是输出两行,第一行为aaa,第二行为null。
分享到:
评论

相关推荐

Global site tag (gtag.js) - Google Analytics