One of the functions of the init() method is to start the processes on the machines specified in the process-config. When the user starts his parallel application on a specific machine. The process started by the user becomes what is called the initiator of the parallel application. This initiator starts the process on different machines specified in the process-config file. The first process in the process-config becomes the master in the group of parallel processes. The other processes become the slaves. With this kind of setup, one can start off the processes from a remote machine itself and run the parallel application. Using this approach, the initiator will not consume any CPU on the machines in which the parallel application is being run. Another advantage of this is that, the initiator can wait for all the other process to finish and then return control back to the user. The user will be able to know when all the applications started have finished execution. This initiator was designed for future extensions to the message passing library. Some functions which may need an external process to do the work can be implemented with the help of the initiator process.