Toward a new approach for sorting extremely large data files in the big data era