The diagram below shows a general view of the architecture of the full MCU video conference solution:
The elements composing the solutions are:
Asterisk or any other SIP Proxy or Registrar for the video phones
SIP Application Server, for example Sailfin or Mobicents Server. It runs the mcuWeb application and handles the SIP signalling part of the service.
mcuWeb which is a SIP Servlet application that implement all the business logic of the service. It handles the SIP signalling, commands the mcu through the XML RPC intergace and provides an administration Web Interface for operating the service
MCU Media Mixer which is in charge of handling all the media processing
When the SIP proxy receives a video call INVITE from a soft phone that is routed to a conference (how to do this is dependent of the specific SIP proxy server), it forwards it to the SIP Application Server. The SIP signalling is handled by the mcuWeb application, establishing the call and joining it to the desired multiconference, creating it if necessary.
The mcuWeb establish the RTP media between the SIP proxy or the video phone and the mcu, interacting with the it via the HTTP XML RPC interface.