It's a small application which runs multiple tasks. A Task is an object which implements Runnable interface. TED manages multiple threads and queues tasks if required.

To communicate (add new task, get information about number of running/scheduled tasks) TED uses HTTP protocol:

• GET /tasks/schedule?class=fully.qualified.class.Name - schedules new task
• GET /tasks/count-scheduled - returns current number of scheduled tasks
• GET /tasks/count-running - returns current number of running tasks

After cloning the repository you can use SBT (Simple Build Tool) to build and run application. In the TED's main directory:

$ sbt run -Dhttp.port=8330

where 8330 is number of TCP/IP port on which TED will be listening.

To run the test go to the project's directory and:

$ sbt test

There are already two examples tasks with fully qualified class names:

• example.Itemize which prints every second one of the 5 popular JVM languages
• example.Calculate which simulates calculations (long loop which does nothing)

For examlpe - to schedule the task Itemize run:

$ wget http://localhost:8330/tasks/schedule?class=example.Itemize

when TED is running. The 5 JVM languages will be printed in TED's console.

During the tests I realized that optimal number of running threads depends on the task type. For Itemize it was few thousand threads and for Calculate only tens (on my i7 machine). Maximal number of running threads is preset to 100. This constant can be modified in app/models/Scheduler.scala source file.