[ns] Bi-Directional VoIP simulator

[Pedro Fortuna]

Hello Lekkie,

I'd like to know the same thing :)
I've been using the Pareto On/Off, but there's no correlation
whatsoever if two flows are set between the same pair of nodes:
http://www.isi.edu/nsnam/ns/doc/node495.html

I guess it shouldnt be too hard to take the source code of the pareto
traffic classes (maybe developing a subclass of Pareto) and develop a
new Traffic Source with bidirectional knowledge... but at the momment
I have too much things in hand to look into that.
If you find or develop anything, please let me know :)

Best Regards,
Pedro Fortuna
INESC Porto

On 8/15/06, lekkie omotayo <lekkie.aydot at gmail.com> wrote:
>
> Hi Pedro,
>
> I tried your sample code for VoIP traffic simulator. it worked fine as a
> uni-directional VoIP simulator but it is not bi-directional.
>
> I need something that can emulate a real life phone conversation and not a
> political debate conversation.
>
> If u ve got any idea, I will appreciate it. Also, some quarters believed CBR
> is better than Pareto distribution for VoIP emulation.
>
> I look forward to ur reply.
>
> Thanks.
>
> Lekkie.
>
>
> > Because, in that particular example, the node is just the sink of the
> > VoIP traffic. It does not send packets in the reverse direction.
>
> > If bidirectional VoIP flows was desired, then you would need an udp0
> > in one direction, and another in the reverse direction.
> > Something like:
> > set udp0 [new Agent/UDP]
> > $ns attach-agent $n(1) $udp0
> > set udp1 [new Agent/UDP]
> > $ns attach-agent $n(2) $udp1
> > $ns connect $udp0 $udp1
>
> > Regards,
> > Pedro Fortuna
> > INESC Porto