[ns] How to avoid movementTrace of mobile nodes ? ( It seems movementTrace OFF does not work)
I want to simulate an ad hoc mobile network with 100 mobile nodes in 1000
seconds. But the problem is it take too much memory space and time to
preform simulation. In order to remove unnecessary info of trace files, I
tried to remove movement trace of mobile nodes with movementTrace OFF in
node-config . After this I found that the trace file has the same result
when movementTrace is either ON or OFF. I mean lines like:
* M 1.0 0 (490.18, 456.19, 0.00), (0.00, 0.00), 0.00
** M 1.0 1 (130.68, 257.04, 0.00), (0.00, 0.00), 0.00 *
* M 1.0 2 (88.83, 394.25, 0.00), (0.00, 0.00), 0.00
M 1.0 3 (236.69, 88.64, 0.00), (0.00, 0.00), 0.00
M 1.0 4 (264.81, 234.03, 0.00), (0.00, 0.00), 0.00
M 1.0 0 (490.18, 456.19, 0.00), (377.91, 111.04), 2.26
M 1.0 1 (130.68, 257.04, 0.00), (186.11, 421.37), 3.98
M 1.0 2 (88.83, 394.25, 0.00), (244.22, 22.03), 3.33
M 1.0 3 (236.69, 88.64, 0.00), (192.80, 465.19), 3.81
M 1.0 4 (264.81, 234.03, 0.00), (103.61, 340.43), 0.95
*So, I changed the relating code in ~ns/common/mobilenode.cc - proc
log_movement() . After having compiled, the new mobilenode.o obejct was
created. I run the silmulation with new created object, but I found that
above lines ( movement trace lines in trace file) didn't change at all. It
seems these lines are coming from the other place. I searched all files that
might be the reason of creation of these lines of trece file, but i did not
find any thing.
void MobileNode::log_movement() {
if (!log_target_) return;
Scheduler s = Scheduler::instance();
sprintf(log_target_-pt_-buffer(),
M %.5f %d (%.2f, %.2f, %.2f), (%.2f, %.2f), %.2f, s.clock(), address_,
X_, Y_, Z_, destX_, destY_, speed_);
log_target_-pt_-dump();
}
The question is where these lines of trace file are coming from?
I would appreciate any help
Amir
[ns] what is the meaning of $self next
Hi, I'm using some test scripts and I found the following code in one of them: $self next What does it mean? It is part of an instproc. There is no function called next in the script. Thank you. Regards, Guillermo
Re: [ns] IPv6 support in NS-2
Check how this is done in mobiwan - these nice guys have implemented IPv6
and Mobile IPv6
http://tagus.inesc-id.pt/~pestrela/ns2/mobility.html
Pedro Vale Estrela
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Tubtim Sanguanwongthong
Sent: sábado, 7 de Outubro de 2006 14:17
To: ns-users@ISI.EDU
Subject: [ns] IPv6 support in NS-2
I've tried to modify NS-2 to support IPv6 beginning with ip.h.
In ip.h,
struct hdr_ip {
_/* common to IPv{4,6} */
ns_addr_t src_;
ns_addr_t dst_;
int ttl_;
}
I'd like to know whether I have to change type of src_ and dst_ to
support 128 bits. ns_addr_t src_; and ns_addr_t dst_; can be used in
both IPv4 and IPv6.
I'm really appreciated in this help (very important).
Thank you in advance.
Tubtim
-
Want to be your own boss? Learn how on Yahoo! Small Business.
Re: [ns] what is the meaning of $self next
http://www.openmash.org/developers/docs/otcl-doc/doc/class.html -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Guillermo Biot Sent: segunda-feira, 9 de Outubro de 2006 11:43 To: Ns Users Subject: [ns] what is the meaning of $self next Hi, I'm using some test scripts and I found the following code in one of them: $self next What does it mean? It is part of an instproc. There is no function called next in the script. Thank you. Regards, Guillermo
Re: [ns] How to change the parameters during the simulation to save thetime
Use this proc:
http://tagus.inesc-id.pt/~pestrela/ns2/contributed_code.html#_Toc147652228
sample code:
global opt
set opt(opt_conv) {
{ d my_delay_ms }
...
}
set opt(my_delay_ms) 100;# default: 100 ms delay
my_getopt $argv
set opt(my_delay_s) [expr $opt(my_delay_ms) * 0.001 ] ;# convert to
seconds
...
$ns make-lan $TCP_BS $TCPr 2mb $opt(my_delay_s) LL Queue/DropTail Mac
...
change the trace name to include the current delay parameter, to create
different trace files
Calling format in bash:
---
for i in 1 10 100 1000; do
ns my_script.tcl -- -d $i
done
Pedro Vale Estrela
http://tagus.inesc-id.pt/~pestrela/ns2/
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Bilal AbuQadous
Sent: sábado, 7 de Outubro de 2006 23:05
To: ns-users@ISI.EDU
Subject: [ns] How to change the parameters during the simulation to save
thetime
Dear ns users,
would you please tell me how I can change the parameters without stopping
the simulation.
as eg.
how I can change the delay in the following command without stopping the
simulation and make that automatic until the end.
$ns make-lan $TCP_BS $TCPr 2mb 0.001ms LL Queue/DropTail Mac
if I will make a test with different delay starting from 0.001, 0.01,
0.01,1,
10, 100, 150 ms, how I could make these changes automatically without
entering new delay parameter by stop the simulation.
I need this thing To save the time and the effort.
thank you in advance
--
AbuQadous, Bilal
Alamiah Electronics (Systems Engineer)
Mobile:+966507174541
Office:+96638610044 ext:772
P.O. Box 9912
Dammam 31423
KSA
[ns] Problem: CBQ with wireless ad hoc network
hi i would be greatly thank full to you if you help me how can i use Class based queue in wireless ad hoc network. i know mobile nodes has interface queue, and default is Queue/DropTail/PriQueue , how i can change it to CBQ/WRR. also, in cbq we have to make link between two nodes and then install priority classes into those links, but in Wireless network we have radio medium between two or more nodes, how can we put priority classes in this case. i m waiting for quick response. thank you bye - Yahoo! Messenger with Voice. Make PC-to-Phone Calls to the US (and 30+ countries) for 2¢/min or less.
[ns] [Fwd: Backward Compatibility mode!]
Original Message Subject: Backward Compatibility mode! From:[EMAIL PROTECTED] Date:Fri, October 6, 2006 1:15 pm To: Ns-users@ISI.EDU -- Hey Guys, I'm a graduate student working on my master's project which is on WSN and i'm just starting my simulations with NS and the computer that we work on has more than one user using it, so we are not allowed to change anything in the BASH files, so when I started simulating with ns it gives me these errors,there are like a bunch of them i've already installed ns, and changed the paths but when try to use th ecommand ls after i changed the paths, it gives me the error -bash: ns: command not found And then when i tried ./ns in a particular directory it started the ns, but when i try running a program, it gives me the error : warning: using backward compatibility mode error when calling class OldSim: filename.tcl can somebody please help with what i'm doing wrong or what i need to do inorder to get this working. Thanks a lot Fatema.
Re: [ns] sending packets
Inside my recv() function I do like this without problems: // clone the current packet and change some fields Packet* new_p = clone_pkt(p); hdr_ip*new_iph = hdr_ip::access(new_p); new_iph-daddr() = ...; new_iph-saddr() = addr(); // change fields on the current packet iph-daddr() = id2iaddr(th-b_id); iph-saddr() = addr(); // send original packet somewhere send(p, 0); // send cloned packet elsewhere send(new_p, 0); ... -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Larry Brigman Sent: segunda-feira, 9 de Outubro de 2006 19:14 To: ns-users@ISI.EDU Subject: [ns] sending packets I am working on a queuing model. There are times that I need to send more than one packet in the recv function. Because of the way NS models things it looks like after a single send or target-recv(p,h) I cannot send again without an error. Is there a way to write a recv function that is allowed to send more than once? Also when I am rescheduling my class to deque packets, about the time I need to deque two or more packets NS detects that time has gone backwards! I am running this on FC4 using 2.29 version of NS. Thanks for any help provided. Larry
Re: [ns] How to compute collision rate on wireless network ?
Hi there, instead of first writing the file and then processing it with awk, pipe the ns output directly through awk so only the lines you are interested in are ever written to disk: set file [open | awk -f myfilter.awk short_tracefile.tr w]; myfilter.awk should contain your awk commands. You can also put the commands directly into the .tcl script, just be sure to escape all quote marks as needed. You can also use any other command, eg. grep or a concatenation of several commands (just pipe one into the next) to pre-process your file. This is the most efficient way to save time and disk space, the only drawback is that you can't later on decide that you want to look at other data since only the filtered data is written. good luck, Martina Umlauft[EMAIL PROTECTED] - http://frauenweb.at/~tina ns user schrieb: hi all, i'm using ns 2.29 to simulate a large wireless topologies (more than 100 nodes). the goal of my simulations is to compute collisions rate. so i 've enabled MAC trace and i'm using awk to filter only lines where there is collision (if $1 = 'D' $5 = 'MAC'). but the problem is that the simulation take a lot of time ( more than 1 hour) because of the trace file that exceed 60 Mo in some cases. So i'm looking if there is not an other method to calculate the collision rate please help me it's urgent ___ Découvrez un nouveau moyen de poser toutes vos questions quel que soit le sujet ! Yahoo! Questions/Réponses pour partager vos connaissances, vos opinions et vos expériences. http://fr.answers.yahoo.com
[ns] Re : How to compute collision rate on wireless network ?
hi Martina; but what i want to do is to compute the following value : number of collided packets / number of sent packet. So i need the two values and even like this i think it will take a lot of time. best regards. - Message d'origine De : Martina Umlauft [EMAIL PROTECTED] À : ns user [EMAIL PROTECTED] Cc : ns-users@ISI.EDU Envoyé le : Lundi, 9 Octobre 2006, 1h00mn 28s Objet : Re: [ns] How to compute collision rate on wireless network ? Hi there, instead of first writing the file and then processing it with awk, pipe the ns output directly through awk so only the lines you are interested in are ever written to disk: set file [open | awk -f myfilter.awk short_tracefile.tr w]; myfilter.awk should contain your awk commands. You can also put the commands directly into the .tcl script, just be sure to escape all quote marks as needed. You can also use any other command, eg. grep or a concatenation of several commands (just pipe one into the next) to pre-process your file. This is the most efficient way to save time and disk space, the only drawback is that you can't later on decide that you want to look at other data since only the filtered data is written. good luck, Martina Umlauft[EMAIL PROTECTED] - http://frauenweb.at/~tina ns user schrieb: hi all, i'm using ns 2.29 to simulate a large wireless topologies (more than 100 nodes). the goal of my simulations is to compute collisions rate. so i 've enabled MAC trace and i'm using awk to filter only lines where there is collision (if $1 = 'D' $5 = 'MAC'). but the problem is that the simulation take a lot of time ( more than 1 hour) because of the trace file that exceed 60 Mo in some cases. So i'm looking if there is not an other method to calculate the collision rate please help me it's urgent ___ Découvrez un nouveau moyen de poser toutes vos questions quel que soit le sujet ! Yahoo! Questions/Réponses pour partager vos connaissances, vos opinions et vos expériences. http://fr.answers.yahoo.com ___ Découvrez un nouveau moyen de poser toutes vos questions quel que soit le sujet ! Yahoo! Questions/Réponses pour partager vos connaissances, vos opinions et vos expériences. http://fr.answers.yahoo.com
Re: [ns] sending packets
On 10/9/06, Pedro Vale Estrela [EMAIL PROTECTED] wrote:
Inside my recv() function I do like this without problems:
// clone the current packet and change some fields
Packet* new_p = clone_pkt(p);
hdr_ip*new_iph = hdr_ip::access(new_p);
new_iph-daddr() = ...;
new_iph-saddr() = addr();
// change fields on the current packet
iph-daddr() = id2iaddr(th-b_id);
iph-saddr() = addr();
// send original packet somewhere
send(p, 0);
// send cloned packet elsewhere
send(new_p, 0);
That doesn't seem to work for me when my class inherits from the queue class.
In my recv function I am doing:
while ((pkt = deque()) != NULL) {
send(pkt,0);
}
I get this error:
Scheduler: attempt to schedule an event with a NULL handler. Don't DO that.
If I change it to use the handler that is inherited from the queue class as
while ((pkt = deque()) != NULL) {
send(pkt,qh_);
}
I get an error like:
Scheduler: Event UID not valid!
If I only do process one packet each time recv is called but always
schedule() with a small
delay, I get NS time backwards error.
I need to schedule every 5-10ms even if no packets are arriving as the
object may have
held packets because of the queuing functions.
[ns] Can nix-vector routing be applied for simulating wireless ad hoc networks?
Hi all, I want to reduce simulation time and memory usage by techniques introduced in Tips and Statistical Data for Running Large Simulations in NS. I want to know whether nix-vector routing can be applied for simulating wireless ad hoc networks? If yes, Is any setting or command required in tcl script in addition to $ns set-nix-routing ? I would appricate any help Amir
Re: [ns] sending packets
No, I need to send different p during the same recv function. (ie I need to deque multiple packets) On 10/9/06, Raghu-kisore Neelisetti [EMAIL PROTECTED] wrote: Hi, Iam not sure If i understood you right. You cannot send the same 'p' over and over again (If this what u meant). The way to avoid the problem is target-recv(p,h) //Now make a copy of the p that you sent before and then target-recv(p,h) Iam sorry if i misunderstood you. raghu. Larry Brigman [EMAIL PROTECTED] 10/09/06 1:14 PM I am working on a queuing model. There are times that I need to send more than one packet in the recv function. Because of the way NS models things it looks like after a single send or target-recv(p,h) I cannot send again without an error. Is there a way to write a recv function that is allowed to send more than once? Also when I am rescheduling my class to deque packets, about the time I need to deque two or more packets NS detects that time has gone backwards! I am running this on FC4 using 2.29 version of NS. Thanks for any help provided. Larry
