Hi Jannick,

I recall somebody having asked a similar question. I think the answer
then was that real networks are never that simple and too many protocols
are designed with a bad idea of the real world. Therefore the designers
of TOSSIM decided not to support such a "dummy mode" to force people to
think about the radio environment.

Personally I think a "dummy mode" would also be good for testing and
debugging programs. In such a mode you could design scenarios with a
given characteristic and then check whether your program behaves the way
you intended it to. This would allow, for instance, unit tests.

To the list: Is it correct that a "dummy mode" currently is not
supported for the reasons I mentioned above?

Also to the list: While we probably all understand that running
simulations for protocol evaluation in a dummy mode is non-sense, who
else would be in favor of a "dummy mode" for program testing and debugging?

Cheers,
Urs


Jannick Bitsch schrieb:
> Hi
> 
>  
> 
> I'm following a course on protocols, where my group have to implement a
> simple multi-hop routing protocol on the tiny-os platform (ver 2). We know
> practically nothing about the workings of radio communication and would
> therefore, at least for the time being, prefer to keep the simulation setup
> free of all the quirks of radio communication (packet loss due to noise,
> asymmetric links etc). In essence, to dumb the simulation down to where the
> only "physical" limitation, is the range of the radio signal.
> 
>  
> 
> What we have been looking at so far is:
> 
>  
> 
> Turning off the simulation of noise, if such thing even makes sense. Our
> tests have been based on the example in the tossim tutorial, and if we
> remove the part of the python script that loads and builds a noise model, no
> packets will ever be received. It seems that the network simulation wont
> work unless you supply at least 100 noise-data-points and build a noise
> model. Can you just supply no-noise data-points?
> 
>  
> 
> Building the topology/gain info by using example 3 from
> file:///opt/tinyos-2.x/doc/html/tutorial/usc-topologies.html on our
> topologies.
> 
>  
> 
> We would very much appreciate some input on how we can accomplish this goal
> / what kind of setup would seem suitable for our usage scenario.
> 
>  
> 
> Regards
> 
> Jannick
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to