Reference no: EM133188701
A Simplified Master/Worker Framework
The application background
Volunteer computing is to harness the donated computing cycles and temporary storage from millions of internet volunteers for a computation task, which is too big to be processed by a single computer in a reasonable amount of time. The following web page outlines volunteer computing.
The most successful volunteer computing framework is Berkeley Open Infrastructure for Network Computing (BOINC). BOINC is client/server architecture. The server (termed as master) creates a list of compute-tasks. Volunteer clients (termed as worker in BOINC) access the master to download available compute-tasks. Workers perform the compute-tasks and upload the results to the master. Optionally workers may be awarded credits by the master for their contribution to computing cycles and temporary storage. Volunteer computing by using BOINC framework has been applied to many scientific projects such as SETI@home, Einstein@Home, and IBM World Community Grid. The following web pages introduce volunteer computing and application projects.
In this assignment, you are to develop a simplified master/worker framework. The Java networking models and components that you have practised from Week 1 to Week 4 of this unit are used to develop such a simplified framework. The models and components are the client/server model, Java TCP streaming, multi-threading and object serialization. You will need to review the models and components and practise relevant lab projects of the weeks before you start this assignment.
Note:
1. A separate document for demonstration of the system functions is available on the unit Moodle site. You will need to use the demonstration document as a part of the assignment specification. You will need to ensure that you fully understand the scenario and project specification before developing the project; you will also need to ensure the developed system fulfilling the functional requirements as shown in the demonstration document.
2. A file transfer sample code and instruction document are on the unit Moodle site. It helps for transferring tasks from the Master to the Worker in the framework.
3. The algorithms of the 3 sample tasks are provided on the unit Moodle for the freamwork.
Part 1: Java TCP Streaming, Multi-threading and Object Serialization Programming
The framework consists of a Master (i.e. server), a number of Workers (i.e. clients) and a Class Repository in the Master. The framework is depicted in the following diagram. The framework is a generic computing architecture because the Master and Workers just need to know/follow the interaction contract only so that they can interact with each other via the framework. The specification of the interaction contract is as follows.
![585_Java TCP Streaming.jpg](https://secure.expertsmind.com/CMSImages/585_Java TCP Streaming.jpg)
1. The interaction contract
The interaction contract between a Worker and the Master consists of:
a. The Task interface defines two standard methods that every compute-task must implements.
package Contract; public interface Task {
public void executeTask(); public Object getResult();
}
b. The TaskList class is a container that holds the titles and the class names of available compute-tasks.
package Contract;
import java.io.Serializable;
public class TaskList implements Serializable{ private String AvailableTasks[];
private String TaskClassName[];
public String[] getAvailableTasks() { return AvailableTasks;
}
public void setAvailableTasks(String[] AvailableTasks) { this.AvailableTasks = AvailableTasks;
}
public String[] getTaskClassName() { return TaskClassName;
}
public void setTaskClassName(String[] TaskClassName) { this.TaskClassName = TaskClassName;
}
}
c. The TaskObject class is a container that holds a particular compute-task object, its ID and credit.
package Contract;
import java.io.Serializable;
public class TaskObject implements Serializable{ private Integer TaskID=0;
private Integer Credit=0; private Task TObject=null; public TaskObject() {
}
public Integer getTaskID() { return TaskID;
}
public void setTaskID(Integer TaskID) { this.TaskID = TaskID;
}
public Integer getCredit() { return Credit;
}
public void setCredit(Integer Credit) { this.Credit = Credit;
}
public Task getTObject() { return TObject;
}
public void setTObject(Task TObject) { this.TObject = TObject;
}
}
The above interface and classes form a complete interaction contract between the Master and Workers.
2. The compute-task
The Master has a class repository that holds the Java classes of available compute-tasks. A compute-task must implement the Task interface. Executing the executeTask() method will perform the task and set the result. Calling the getResult() method will return the result. A compute-task must implement java.io.Serializable interface so that the compute-task can be transferred between the Master and a worker over the network. The structure of a compute-class is as follows.
public class CalculatePi implements Task, Serializable{
...... @Override
public void executeTask() {
//The implementation of method
......
}
@Override
public Object getResult() {
//The implementation of method
......
}
//may have other methods
......
}
3. The interaction workflow of the framework
A Worker provides a user frontend to access the remote Master. A Worker just needs to follow the interaction contract only for volunteer computing. The interaction workflow between Worker and Mater includes:
a. Workers connect to the Master
b. Workers download the available task list from the Master
c. Workers download some tasks from the Master
d. Workers perform the tasks
e. Workers upload computing results to the Master.
The demonstration of interaction workflow between Workers and the Mater is in the demonstration document on the unit Moodle site.
4. The implementation
To complete this assignment, you need to implement such a framework and integrate the Calculate Pi, Calculate Primes and Calculate the Greatest Common Divisor tasks into this framework. The algorithms of the tasks are given on the unit web site. The Master must be multi-threaded and follow the ‘thread-per-connection' architecture (reference Week-4 contents). The communication between the Master and Workers must use TCP streaming by using Java TCP API Socket and ServerSocket as described in Week-2 contents of this unit. Note: use of any other networking protocols will incur no marks to be awarded for this part.
To implement the framework, you need to implement the following Java classes:
a. A Java application to implement the Worker; a graphic user interface is required;
b. A Java application to implement the Master; and
c. A number of Java classes to implement the processing threads when necessary.
d. Note: to demonstrate your competence in concurrency programming, you will need to make an analysis on when concurrency is needed. Marks will be deducted if concurrency is not identified or necessary multithreading is not used.
e. A number of Java classes to implement Calculate Pi, Calculate Primes and Calculate the Greatest Common Divisor tasks.
Note: to simulate Master and Worker interaction, you don't have to run them on two physical machines. Instead, they can be run on two JVMs (Java Virtual Machines) on a single physical machine. As a result, the name of the Master machine can be ‘localhost'.
Part 2: Program use and test instruction
After the implementation of the framework, prepare an end user' instruction about how to use your software. The instruction should cover all aspects of the framework as detailed in the Part 2 of marking criteria below.
Submission
You need to provide the following files in your submission.
1. Files of Java source code of the Master, the Worker and the processing threads and the compute-tasks. The in-line comments on the data structure and program structure in the programs are required. The source code files must be able to be compiled by the standard JDK (Java Development Kit) or NetBeans IDE from Oracle.
2. The compiled Java class files of the source code. The Java classes must be runnable on the standard Java Runtime Environment (JRE) from Oracle.
Note: an easy way to provide the source code and executables is to submit them in a NetBeans project.
3. A Microsoft Word document to address the issues as specified in Part 2 above.
All the required files must be compressed into a zip file for submission. You must submit your assignment via the unit web site. Any hardcopy or email submission will not be accepted. After the marked assignments are returned, any late submissions will not be accepted.
Attachment:- Simplified Master.rar