[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: problem with running snoop.tcl with ns



Hi,

As in my previous mail, I had indicated that when you feed snoop.tcl to ns
with different error rates, you don't get different throughput but you get
the same throughput and similar "out" and "tcp.tr" files along with almost
similar values for variables like "t_seqno_", "maxseq_" etc. in the end of
tcp.tr file.

I tried to debug what is going on and found that ErrorModel::corrupt() in
errmodel.cc is never called when I feed snoop.tcl to ns. I think for
getting corruption, this is needed, isn't it? If yes, why this is
happening? And if not then what else is the reason for not getting   
different throughput with different error rates (like 10%, 1%, with snoop
or without snoop enabled) 

I would like to ask those ns users who have run snoop.tcl that, do you
find any difference while running snoop.tcl with different error rates
(like 1%, 10% etc.) and also with "-ll Snoop" and without it.

I have *SERIOUS DOUBTS* about generation of error while running it. I want
to check whether that is common with you all or something specific with
me. Please check this. I have ns version ns2.1b2 released on Jan 29th. 

Thanks,

Miten.

----------------------------------------------------------------------
Miten N. Mehta                                 	Graduate Research Assistant 
3902, Collegemain, #908				Computer Science
Bryan, 						Texas A & M University
Texas-77801.					College Station, TX-77843

Tel.#(409) 846 1451                            	(409) 845 5007 
			E-mail# [email protected]
            Home Page : http://www.cs.tamu.edu/people/mmehta

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

	"UNTIL YOUR DREAMS COME TRUE, KEEP ON WORKING FOR THAT"

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