Network Traffic Flow learning and Simulation

[Mayank-Bhatnagar <mayank () ncb ernet in>]

hi all,

Well one of the things that has always interested as part of our network
experiments are generating Network Traffic Flow.

Statistical and Anomaly based Intrusion Detection Systems require a
"clear" data during their modelling phase.

Apart from deploying the network sniffer (data capturer) and getting
traffic, there are various simulating tools that would simulate the
network. One such tool is ne (http://www.isi.edu/nsnam/ns/).

ns-2 has some tips for generating large simulation, but might
not be catering to recreating large simulation of the actual network.

This is in fact what we are looking for.

More specifically, we would like to know, are there some tools available
that would perhaps be an integration of sniffer and "learning" ability, to be
plugged in to the live network, understand its behaviour, and then can be
unlpugged, and be customised to generate similar kind of traffic
(high speed traffic) for various statistical IDS tests & experiments.

If yes, it would be great to get some pointers/papers. It would be nice if
someone among you, pointed out any experience in coming up with these
learning testing tools.


Thanks & Regards,
Mayank Bhatnagar
mayank () ncb ernet in
68 Electronics City ,
NCST, Bangalore-560100.
080-28523300/28520259-1200




---------------------------------------------------------------------------

---------------------------------------------------------------------------