The concurrent way to do it is to launch all requests at once, then switch among them as the responses come back. Through ... make use of all those cores, if possible. For parallelism, Python ...