In the multiprocessing module, the Pool class is mainly used to implement a pool of processes, each of which will carry out tasks submitted to a Pool object. Generally, the Pool class is more convenient than the Process class, especially if the results returned from your concurrent application should be ordered.

Specifically, we have seen that the order of completion for different items in a list is considerably likely to change when put through a function concurrently as the program runs over and over again. This leads to difficulty when reordering the outputs of the program with respect to the order of the inputs that produced them. One possible solution to this is to create tuples of processes and their outputs, and to sort them by process ID.

This problem is addressed by the Pool class: the Pool.map() and Pool.apply() methods follow the convention of Python's traditional map() and apply() methods, ensuring that the returned values are ordered in the same way that the input is. These methods, however, block the main program until a process has finished processing. The Pool class, therefore, also has the map_async() and apply_async() functions to better assist concurrency and parallelism.