[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ns] Changing TCP window (and sat parameter qlim) doesn't affect simulation results (?!)
Hi all,
I have made these changes to the script sat-repeater.tcl that comes with NS.
set opt(bw_up) 2Mb; ---> set opt(bw_up) 10Mb;
set tcp1 [$ns create-connection TCP $n2 TCPSink $n3 0] ---> set tcp0 [$ns
create-connection TCP/Newreno $n2 TCPSink $n3 0]
$ns at 7.0 "$ftp1 produce 100" ---> $ns at 7.0 "$ftp1 start"
I have also written a simple script in AWK to calculate the throughput.
According to this script the throughput is very low.
(few hundreds of kb/s) Is it correct or is the script wrong?
Probably the script is wrong but
I really don't understand why changing TCP window, $tcp1 set window_ ....
all you want, I don't see any difference in the results. Please, help me.
I have also modified the parameter qlim. As reported in the manual, it's
the length of the interface queue, in packet. If I have understood well the
figure 17.3 of the manual, this queue is between Mac and LL, so these
packets are in other words frames. However I haven't seen particular
changes in the simulation results. Is it correct? I have made a mistake
..., I haven't understood something ...
Please tell me something
And if the throughput is low, what can I do to increase it?
Below you can find the AWK script.
Any help is appreciated. Thank you in advance,
Giuseppe Tringali
OS: Linux Red Hat 6.2
NS: ns-allinone-2.1b6
AWK: GNU Awk 3.0.4
BEGIN {slot = 0.5; bytes_counter = 0; t_sampling = 0.5;}
{
action = $1;
time = $2;
node_1 = $3;
node_2 = $4;
size = $6;
type = $5;
if ( (action == "r") && (node_2 == "2") && (type == "tcp") ) {
if ( time <= t_sampling) bytes_counter = bytes_counter + size;
if ( time > t_sampling) {
th = ( (bytes_counter * 8) / (1000000 *
slot) );
printf("%f %f\n", t_sampling, th);
bytes_counter = size;
t_sampling = t_sampling + slot;
}
}
}
END { }