This allows a derived (concrete) implementation to share information (e.g. sample rate, buffer size) between the audio backend side of things and the port management side of things.