Re: [ns] XGraph

[Tarik Alj <aljtarik@ozias.inrs-telecom.uquebec.ca>]

> Thanx Tarik for replying , I had a better solution for this problem ( I see it
> is useful and easier than awk that I found very confusing), I used the vi 
editor

as you wish... but then why not Sed? The thing with awk is that it handles files 
very nicely and it often takes very simple scripts to get just any kind of data 
that needs to be exctracted of the traces, like end to end delay or inter-packet 
jitter (diff between inter-emission times and inter-arrival times).

> commands to adjust the file, and it was so easy
> Also I think that the xgraph is a very useful plotting program to use, but I
> have a comment, that why only two columns at each time??

Like Lloyd said that's pretty much the standard way of doing it.

> I think it will be more useful if it had multiple columns and can be plotted 
at
> the same time (you know, like MS Excel)

Like Lloyd said it can take multiple files as an argument, or you can separate 
multiple data set by specifying a title :

"set 1"
1 34
2 35
...

"set 2"
1 56
2 78
...

that's all in the man page that is included with the distribution of xgraph that 
comes with allinone.

and as far as Excel is concerned, it is MS; need I say more?

> Right now I am in the plotting part of the different simulations, but I have a
> simple question for you...
> How can I define the step time of the simulation?

what do you mean by step? If you are using trace-all it simply records ALL the 
events, wich, yes, can lead to pretty huge trace files. There are many ways to
trace stuff during a simulation :

- use monitors
- use "record" procedures (check the tutorial)
- build your own tracing procedures (there are Trace objects both at the C++ and 
the OTcl level you can play with)
- have *fun* with awk by piping the trace-all output throught awk and filter or 
directly calculate what you need from it.

The latter one can rather slow your simulation as you would be running awk and 
ns at the same time.

May I suggest you get your hands on the tutorial and the manual and browse 
throught tcl/ex or tcl/test for examples?
  
> The problem is I run my simulation with 20 nodes and for 200 seconds,
> the stpe is already specified (till now I don't know where), it creates
> a HUGE data file....I need to increase the step size or increase the
> period that the simulator gets data with
> 
> Please advice
> 
> 
> Mohamed Ahmed
> 

gluck.

Tarik