Control threads in AMF

AMF allows you to schedule and manage the lifespan of threads. The following explains how to create and use a scheduler in AMF.

info

Use of control threads in AMF is specific to Java or Scala applications.

In AMF, common uses of thread control include:

• Addressing memory constraints
• Controlling thread creation
• Inserting time intervals between AMF executions

Controlling thread generation and lifespan

AMF allows you to control thread generation and lifespan by customizing the ExecutionEnvironment object. To customize the ExecutionEnvironment object you need to:

1. Create a java.util.concurrent.ScheduledExecutorService
2. Create an amf.client.execution.ExecutionEnvironment with the new scheduler
3. Plug the ExecutionEnvironment into AMF initialization or in resolution stage

Create a ScheduledExecutorService

To create a thread pool that can schedule commands to run after a given delay, or to execute them periodically, create a ScheduledExecutorService:

ScheduledExecutorService scheduler = Executors.newScheduledThreadPool(30, new ThreadFactory() {
            int i = 0;

            @Override
            public Thread newThread(Runnable r) {
                i = i + 1;
                final Thread thread = new Thread(r);
                thread.setName("AMF-" + i);
                return thread;
            }
        });

This example scheduler keeps 30 threads in the pool (even if they are idle), and creates a specific ThreadFactory that sets the thread name when creating a new thread.

Scheduler API methods

The scheduler exposes the following methods:

• scheduled()
  • Schedule a single task execution after a fixed delay
• scheduleAtFixedRate()
  • Run a task periodically after a fixed delay
• shutdown()
  • Stop the ExecutorService from accepting new tasks and shut down after all running threads finish their current work
• shutdownNow()
  • Return a list of tasks that are waiting to be processed, and destroy the ExecutorService immediately
  • All running threads may not be stopped at the same time

Refer to the Java documentation for more information about the ScheduledExecutorService

Create and add the ExecutionEnvironment in AMF

When initialized, AMF creates a default amf.client.environment.Environment. You can customize this environment by passing an amf.client.execution.ExecutionEnvironment object.

The ExecutionEnvironment is responsible for thread creation, and it can receive a ScheduledExecutorService to use when creating threads.

You can plug your customized ExecutionEnvironment into AMF's initialization, or add it when creating a specific parser.

This customized environment can be plugged into AMF's initialization or when creating a specific parser.

Example code for creating a scheduler

The following example code shows the creation of a scheduler, and its usage throughout AMF:

Code extracted from the examples GitHub repository.