[Battlemesh] ns-3 - Network Simulator

[fboehm]

Am 16.11.15 um 07:56 schrieb Henning Rogge:
> On Mon, Nov 16, 2015 at 4:47 PM, fboehm <fboehm at aon.at> wrote:
>> Interesting.
>>
>> A simulator in emulation mode takes hundreds to thousand times longer to
>> deliver packets to other (maybe emulated or real) network elements. Of
>> course, it's an event-driven piece of complex software.
>>
>> But in this respect a simulator in emulation mode in combination with other
>> (real) network components can no longer deliver meaningful results. Maybe if
>> the emulated components are only source and sinks for generating and
>> consuming real traffic. Or am I wrong?
>
> You can... but the question is how to interpret the results.
>
> Its the same problem with simulation and emulation. If they drift away
> from "reality" too much the results are difficult to apply to real
> networks... at a certain point you might just say "yeah, it happened
> in my test, but its not comparable to a real application/network".
>
> Henning
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
>
I see.

What about interference simulations? Any other useful approaches next to 
event-driven simulation where you can use channel models?

Franz