Thank you all for your remarks, I said in the first mail that simulation with CBR traffic and TCP was working but the problem was that the pacet size of CBR was ignored and the rate at the destination was not constant.
In fact, the CBR is in the application layer and the TCP is in the transport layer. So the CBR is sending packets (having the CBR packetSize) in a constant rate, than if TCP cannot send packetsin the same rate(depending on Tcp packetSize, etc) , CBR packets will be waiting and that's why we cannot observe a constant bit rate at the destination. So finally, the CBR traffic seems to be a good application to compare UDP and TCP. isn't it ?
