Re: [ns] Simultaneous start

["Kun-chan Lan" <klan@usc.edu>]

it seems that someone have post similar problem like yours (but
I couldn't recall the answer ) . you might want to check the mailing list
archive to see if you can get any luck .

Kun-chan Lan

-----Original Message-----
From: Antonello <digioia.m@tin.it>
To: NS bulletin board <ns-users@mash.CS.Berkeley.EDU>
Date: Sunday, June 04, 2000 2:53 AM
Subject: [ns] Simultaneous start


>Hi,
>I'm trying to evaluate performance of queueing algorithm. Actually I'm
>simulating a simple topology:
>2 flows CBR --> flow 1 send 8192 bytes every 33.33 ms  ($cbr1)
>                             flow 2 send 8192 bytes every 8 ms    ($cbr0)
>
>In Tcl code I wrote: $ns at 0 "$cbr1 start"
>                                  $ns at 0 "$cbr0 start"
>If I swap these two lines of code results are quite different....why?
>I need to show that service curve algorithm performs better for flows that
>need fixed delay bound ; if I use code with cbr1 first the results are
>identical (with a simple DropTail); if I use code with cbr0 first service
>curve algorithm appears better. What is wrong in this approach?
>
>Thanks,
>@nto
>