Re: ns and simplex links in nam files?

[Haobo Yu <haoboy@isi.edu>]

This is because nam only handles duplex links when it parses the trace
file. :( I'll try to add that in the next few days.

- Haobo

On Tue, 7 Jul 1998, Lloyd Wood wrote:

> ns 2.1b2 and recent daily snaphots don't appear to put link
> information in output.nam files for simplex links, although they do
> for duplex links. As a result, mixing simplex and duplex links results
> in only the duplex links and their nodes being drawn by nam; if you
> have mixed the two in your network you will end up with lots of
> disconnected sections. 
> 
> If you run a simple script setting up a duplex-link between 0 and 1
> and sending traffic from 0 to 1, change that to simplex-link and run
> it again, and compare the output files, you'll see that the lines
> beginning 'l' specifying the links are missing from the simplex-link
> out.nam file, and that nam has nothing to show you when the simulation
> is played back.
> 
> Are there any obvious workarounds, fixes, or reasons for this that I
> don't know about? Since the networks I'm modelling at present are static,
> running the fully-duplex equivalent simulation to get the necessary
> l -t * etc
> lines, which I then cut-and-paste into the simplex out.nam, strikes me
> as an immediate workaround for setting up the initial network for very
> simple cases. I'm not sure about more complex cases where links are
> set to fail under certain conditions, etc. 
> 
> thanks,
>   
> L.
> 
> (btw, nam 1.0a4 is coping with the equivalent fully-duplex topologies
> I'm throwing at it extremely well - I couldn't see simulated Manhattan
> meshes, cylindrical meshes or fully-fourway-connected toroidal meshes
> at all with 0.8, but 1.0a4 handles these easily and clearly with
> autolayout - although links cross, of course. Now, if only I could
> rotate the network in 3D or have shading of crossing links to give
> some separation and depth...)
> 
> <http://www.ee.surrey.ac.uk/Personal/L.Wood/>PGP<L.Wood@surrey.ac.uk>
> 
> 
> 
>