Re: [ns] ns scalability

[Will Shaver <will@digitalfountain.com>]

Florentina Popovici wrote:
> 
> Hey all,
> 
> I plan to use ns for simulating an application-level protocol over TCP. The
> network topology that I will use has a number of nodes anywhere between 100
> and 1000.
> 
> Does any of you have experience in using ns with a large network? How well
> does ns scale? What are the elements that can be a limit in
> my simulation (e.g. number of events?).
> 
> Thank you,
> Florentina
> 
> ________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

The thing that slows my simulations down when I have that may nodes is
not the events themselves, but writing the trace files to a disk.
Minimize the number of tracefiles used to exactly what you need, and if
possible write your own output files instead of using the built in trace
files. Otherwise, I've done sim's with over 100 nodes and not had a
problem.