In this site, it was mentioned that the server specs to use is a dual intel
xeon with at least 2.0 GHz and 1GB of RAM and 0GB of HDD. Does anyone have any
idea on how much disk space or memory a 1-to-1 multimedia (a/v) chat has? Is
there a way to measure say, given the number of bandwidth, this is how much
memory I need??

Take a peek at Brian Lesser's book Programming Flash Communication Server. Reference for all that is in the back.