O'Reilly logo

Java Threads, Second Edition by Henry Wong, Scott Oaks

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

wait(), notify(), and notifyAll()

What happens when there is more than one thread waiting for the notification? Which thread actually gets the notification when notify() is called? The answer is that it depends: the Java specification doesn’t define which thread gets notified. Which thread actually receives the notification varies based on several factors, including the implementation of the Java virtual machine and scheduling and timing issues during the execution of the program. There is no way to determine, even on a single platform, which of multiple threads receives the notification.

There is another method of the Object class that assists us when multiple threads are waiting for a condition:

void notifyAll()

Notifies all the threads waiting on the object that the condition has occurred. This is a method of the Object class and must be called from within a synchronized method or block.

The Object class also provides the notifyAll() method, which helps us in those cases where the program cannot be designed to allow any arbitrary thread to receive the notification. This method is similar to the notify() method, except that all of the threads that are waiting on the object will be notified instead of a single arbitrary thread. Just like the notify() method, the notifyAll() method does not let us decide which threads get notification: they all get notified. By having all the threads receive notification, it is now possible for us to work out a mechanism for the threads to choose among themselves which thread should continue and which thread(s) should call the wait() method again.

Why would you want to wake up all of the threads? There are a few possible reasons, one of which is if there is more than one condition to wait for. Since we cannot control which thread gets the notification, it is entirely possible that a notification wakes up a thread that is waiting for an entirely different condition. By waking up all the waiting threads, we can design the program so that the threads decide among themselves which should execute next.

Another reason is the case where the notification can satisfy multiple waiting threads. Let’s examine a case where we need such control:

public class ResourceThrottle {
    private int resourcecount = 0;
    private int resourcemax = 1;

    public ResourceThrottle (int max) {
        resourcecount = 0;
        resourcemax = max;
    }

    public synchronized void getResource (int numberof) {
        while (true) {
            if ((resourcecount + numberof) <= resourcemax) {
                resourcecount += numberof;
                break;
            }
            try {
                wait();
            } catch (Exception e) {}
        }
    }

    public synchronized void freeResource (int numberof) {
        resourcecount -= numberof;
        notifyAll();
    }
}

We are defining a new class called the ResourceThrottle class. This class provides two methods, getResource() and freeResource(). Both of these methods take a single parameter that specifies how many resources to grab or release. The maximum number of resources available is defined by the constructor of the ResourceThrottle class. This class is similar to our BusyFlag class, in that our getResource() method would have to wait if the number of requested resources is not available. The freeResource() method also has to call the notify() method so that the waiting threads can get notification when more resources are available.

The difference in this case is that we are calling the notifyAll() method instead of the notify() method. There are two reasons for this:

  • It is entirely possible for the system to wake up a thread that needs more resources than are available, even with the resources that have just been freed. If we had used the notify() method, another thread that could be satisfied with the current amount of resources would not get the chance to grab those resources because the system picked the wrong thread to wake up.

  • It is possible to satisfy more than one thread with the number of resources we have just freed. As an example, if we free ten resources, we can then let four other threads grab three, four, one, and two resources, respectively. There is not a one-to-one ratio between the number of threads freeing resources and the number of threads grabbing resources.

By notifying all the threads, we solve these two problems with little work. However, all we have accomplished is to simulate a targeted notification scheme. We are not really controlling which threads wake up; instead, we are controlling which thread takes control after they all get notification. This can be very inefficient if there are many threads waiting to get notification, because many wake up only to see that the condition is still unsatisfied, and they must wait again.

If we really need to control which thread gets the notification, we could also implement an array of objects whose sole purpose is to act as a waiting point for threads and who are targets of notification of conditions. This means that each thread waits on a different object in the array. By having the thread that calls the notify() method decide which thread should receive notification, we remove the overhead of many threads waking up only to go back to a wait state moments later. The disadvantage of using an array of objects is, of course, that we will lock on different objects. This acquisition of many locks could lead to confusion or, even worse, deadlock. It is also more complicated to accomplish; we may even have to write a new class just to help with notification targeting:

public class TargetNotify {
    private Object Targets[] = null;

    public TargetNotify (int numberOfTargets) {
        Targets = new Object[numberOfTargets];

        for (int i = 0; i < numberOfTargets; i++) {
            Targets[i] = new Object();
        }
    }

    public void wait (int targetNumber) {
        synchronized (Targets[targetNumber]) {
            try {
                Targets[targetNumber].wait();
            } catch (Exception e) {}
        }
    }

    public void notify (int targetNumber) {
        synchronized (Targets[targetNumber]) {
            Targets[targetNumber].notify();
        }
    }
}

The concept is simple: in our TargetNotify class, we are using an array of objects for the sole purpose of using the wait and notify mechanism. Instead of having all the threads wait on the this object, we choose an object to wait on. (This is potentially confusing: we are not overriding the wait() method of the Object class here since we’ve provided a unique signature.) Later, when we decide which threads should wake up, we can target the notification since the threads are waiting on different objects.

Whether the efficiency of a targeted notification scheme outweighs the extra complexity is the decision of the program designer. In other words, both techniques have their drawbacks, and we leave it up to the implementors to decide which mechanism is best.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required