Distributed word counting java program

Hi, Im trying to implement a program to calculate a frequency table for the words that appear in a set of individual text documents using multiple threads in Java. So, for instance given a single file, write a method to compute the frequency of each word that appears in that file. This could be done using the space character as the delimiter for tokenising texts and assume all tokens separated by a space character to be a word. The counting the word frequency distribution in all the files shall be carried out by separate threads. Each thread being implemented as a distinct thread (multithreading). Each thread must first get the list of the names of the files to process from the files containing random text.

Then it shall proceed in phases, in each phase choosing a file at random from those not yet processed by any worker and then processing that file. The program should implement some method to guarantee that no file is processed twice (For each file, only a single thread should ever process that file, of course, without being to tell in advance which thread that will be). Once a thread has computed the word frequency distribution of a single file, it must write the result to a central data structure (shared between all threads), such as an associative array or a hash table named. However, no two threads must be writing to this array/hash table at the same time. Implement this level of thread synchronisation and mutual exclusion using some locking mechanism.

Vary the number of threads from 1 to 100 and measure in each repetition the time to complete the entire task. Plot results in an x y scatter plot, where the x-axis represents the number of threads and the y-axis represents the time taken in milliseconds.