Re: [ns] trouble with error model over wired links
Hi there,
This is another fine post to the ns-users mailing list.
Good subject, good question overall == totally in line with
http://www.catb.org/~esr/faqs/smart-questions.html
the problem is that you are putting the loss-model in all 3 links;
I report exactly the same tracea and NAM problems.
Instead, Try this:
#$ns link-lossmodel $em $n(0) $n(1)
$ns link-lossmodel $em $n(1) $n(2)
#$ns link-lossmodel $em $n(2) $n(3)
pedro vale estrela
http://tagus.inesc-id.pt/~pestrela/ns2/
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Abdul Jabbar
Sent: sexta-feira, 29 de Setembro de 2006 17:45
To: ns-users@ISI.EDU
Subject: [ns] trouble with error model over wired links
Hello,
I am having unusual trouble getting a simple error model to work over
wired links in ns2.29. I have checked the mailing list archives and
could not find a similar problem/solution. Below is my tcl script which
represents a simple network topology of 3 nodes connected in a chain
using duplex-links. I have a CBR/UDP flow between the nodes 0 and 3. The
flow path is 0-1-2-3.
If I don't use any error on the links, everything works fine as is
confirmed by the trace file and the nam visualization.
However, if I include a simple error model and attach it to each of the
three links, the whole simulation goes awry. In the trace file I see
packets going from node 0 to node 1. There is no data flow between node
2 and node 3 and there is a data flow between nodes 2 and 3. Basically
the packets disappear on node 1 and magically reappear at node 2. In
nam visualization, it is even worse...there is just one flow between
nodes 0 and 1thats it.no other data flows. Changing the error
rate does not help.
When I have a more complex topology ( like a mesh topology of 8x8 nodes)
the whole simulation goes so bizarre, that it is even hard to trace what
is going wrong. That's why I am using this simple 3 node topology to
illustrate the problem.
I am hoping that there is some fundamental mistake on my part. Please
point in the right direction.
Thanks,
Abdul.
Here's the TCL script
--
set ns [new Simulator]
set tf [open simout.tr w]
$ns trace-all $tf
set nf [open simout.nam w]
$ns namtrace-all $nf
proc finish {} {
global ns nf tf
$ns flush-trace
close $nf
close $tf
exec nam simout.nam
exit 0
}
for {set i 0} {$i 4} {incr i} {
set n($i) [$ns node]
}
$ns duplex-link $n(0) $n(1) 500Mb 10ms DropTail
$ns duplex-link $n(1) $n(2) 500Mb 10ms DropTail
$ns duplex-link $n(2) $n(3) 500Mb 10ms DropTail
set em [new ErrorModel]
$em unit pkt
$em set rate_ 0.01
$em ranvar [new RandomVariable/Uniform]
$em drop-target [new Agent/Null]
$ns link-lossmodel $em $n(0) $n(1)
$ns link-lossmodel $em $n(1) $n(2)
$ns link-lossmodel $em $n(2) $n(3)
set udp0 [new Agent/UDP]
$ns attach-agent $n(0) $udp0
set cbr0 [new Application/Traffic/CBR]
$cbr0 set packetSize_ 1000
$cbr0 set interval_ 0.005
$cbr0 attach-agent $udp0
set null0 [new Agent/Null]
$ns attach-agent $n(3) $null0
$ns connect $udp0 $null0
$ns at 0.5 $cbr0 start
$ns at 4.5 $cbr0 stop
$ns at 5.0 finish
$ns run
Re: [ns] trouble with error model over wired links
I was able to solve this problem. I would like to thank both Pedro and
Alam for their response. The solution provided by Pedro: that is to
apply the error to just one link does work. However, my intention was to
simulate a network where every link has a certain error rate. The
solution provided by Alam: that is to use a separate errorModel object
for each link works exactly as I intended.
The mistake on my part was that after attaching the errorModel object to
one link, I was trying to attach the same object to a second link.
However, the response of ns was very confusing.
Thanks,
Abdul.
Abdul Jabbar wrote:
Hello,
I am having unusual trouble getting a simple error model to work over
wired links in ns2.29. I have checked the mailing list archives and
could not find a similar problem/solution. Below is my tcl script
which represents a simple network topology of 3 nodes connected in a
chain using duplex-links. I have a CBR/UDP flow between the nodes 0
and 3. The flow path is 0-1-2-3.
If I don't use any error on the links, everything works fine as is
confirmed by the trace file and the nam visualization.
However, if I include a simple error model and attach it to each of
the three links, the whole simulation goes awry. In the trace file I
see packets going from node 0 to node 1. There is no data flow between
node 2 and node 3 and there is a data flow between nodes 2 and 3.
Basically the packets disappear on node 1 and magically reappear at
node 2. In nam visualization, it is even worse...there is just one
flow between nodes 0 and 1thats it.no other data flows.
Changing the error rate does not help.
When I have a more complex topology ( like a mesh topology of 8x8
nodes) the whole simulation goes so bizarre, that it is even hard to
trace what is going wrong. That's why I am using this simple 3 node
topology to illustrate the problem.
I am hoping that there is some fundamental mistake on my part. Please
point in the right direction.
Thanks,
Abdul.
Here's the TCL script
--
set ns [new Simulator]
set tf [open simout.tr w]
$ns trace-all $tf
set nf [open simout.nam w]
$ns namtrace-all $nf
proc finish {} {
global ns nf tf
$ns flush-trace
close $nf
close $tf
exec nam simout.nam
exit 0
}
for {set i 0} {$i 4} {incr i} {
set n($i) [$ns node]
}
$ns duplex-link $n(0) $n(1) 500Mb 10ms DropTail
$ns duplex-link $n(1) $n(2) 500Mb 10ms DropTail
$ns duplex-link $n(2) $n(3) 500Mb 10ms DropTail
set em [new ErrorModel]
$em unit pkt
$em set rate_ 0.01
$em ranvar [new RandomVariable/Uniform]
$em drop-target [new Agent/Null]
$ns link-lossmodel $em $n(0) $n(1)
$ns link-lossmodel $em $n(1) $n(2)
$ns link-lossmodel $em $n(2) $n(3)
set udp0 [new Agent/UDP]
$ns attach-agent $n(0) $udp0
set cbr0 [new Application/Traffic/CBR]
$cbr0 set packetSize_ 1000
$cbr0 set interval_ 0.005
$cbr0 attach-agent $udp0
set null0 [new Agent/Null]
$ns attach-agent $n(3) $null0
$ns connect $udp0 $null0
$ns at 0.5 $cbr0 start
$ns at 4.5 $cbr0 stop
$ns at 5.0 finish
$ns run
