You are so comparing apples with oranges and expecting the same result. In one test you are running Communicator at the remote site, so the traffic going over the VPN is the VoIP traffic. How ...
See more...
You are so comparing apples with oranges and expecting the same result. In one test you are running Communicator at the remote site, so the traffic going over the VPN is the VoIP traffic. How much work do you think Cisco have put into making that VoIP traffix small, using good software based compression codecs to compress the voice. You have good VPN links and without QoS you are getting it to work. Most people need QoS as well. In the second test you are running Communicator in the data center, so the traffic going over the VPN is NOT VoIP traffic but IS virtualised USB of bi-directional audio. Here you are living on the bleeding edge. There are no codecs optimising the data stream out to the remote site. Then you have the Communicator running inside a VM, maybe the CPU timeslicing is also screwing up the software codecs for the voice compression as well on top of your other problem. You will probably find that the VoIP traffic is UDP and handles little bits of slow data fine, where as the virtualised USB is TCP and every time one packet gets delayed it screws the whole thing up. You want to claim that its all the same but its not. This is why the vendors say its not supported. This will change, but these deep problems need to be fixed first. There are a lot of us waiting for this to be sorted, but there is some engineering work to be done. Rod (Sorry to sound like a rant, I need more sleep!) Considering awarding points if this is of use