[ns] reading trace files

[Vietor Davis]

A trace file will generally always be like that. A line is put into it 
every time a packet moves up or down the stack. If you wish to create a 
graph in any reasonable amount of time you must write (or find I 
suppose) tools to parse the traces. They are human readable, but don't 
be fooled into thinking that you can make direct use of them without 
some programmatic help.

-Vietor

Nauman Afzal wrote:

> Thanks that was helpful,
>                                    but I took a look at the trace file 
> and it is 9MB in size and 2800 pages long!!! I mean is it always like 
> that? Thta would take ages to complete a graph, or is there a way 
> round this? What if the simulation time is reduced, shouldn't that help?
> Best,
> Nauman
>
>  
> On 3/4/06, *Vietor Davis* <vietor at trainedmonkeystudios.org 
> <mailto:vietor at trainedmonkeystudios.org>> wrote:
>
>
>     All of the information that you should require in order to understand
>     the trace output can be found in ./trace/cmu- trace.cc
>
>     This is the only source that you can be assured of getting an accurate
>     and current description of what every field in the trace output is. It
>     will remove any ambiguity about what any given trace field is being
>     generated from.
>
>     -Vietor
>
>     Nauman Afzal wrote:
>
>     >Hi,
>     >    I just sucessfully ran wireless.tcl file but am finding it
>     difficult to
>     >read trace files. I mean when I open wireleee.tr
>     <http://wireleee.tr> the info is kind of
>     >cryptic. What do these different fields denote? Any ideas how to
>     read it?
>     >Thanks in advance,
>     >Best,
>     >Nauman
>     >
>     >
>
>