Conferencing, participating in a call where more than two parties can communicate, has always been one of the most sought-of service in telecommunication.
Technology for allowing more than two parties to concurrently talk each other (eg, not in a push-to-talk service) is very much different than the one used for "normal" calls with two participants. In a "normal" call, you only care that media streams goes back and forth from caller to callee. Actually you can have (and often you have) the media streams (via RTP) going directly from caller to callee and back, bypassing FreeSWITCH or any other telecommunication server.
Conferencing, on the other hand, needs a server. The most basic function of a conferencing server is mixing, eg merging, the different media streams that are sent by participants. The resulting merged stream is broadcasted to all participants. So, each participant sends to server its own media stream, and all participants receive from server a media stream...