Skip to content

Charles Pandian

A comprehensive ns-3 Cheat Sheet covering all fundamental skills for network simulation researchers and engineers.

Forum Replies Created

Charles Pandian
Keymaster

Simply you need to add your custom fields in packet header structure of ns-2 and use it in your code.

In packet.h modify the structure hdr_cmn as follows :

struct hdr_cmn {
enum dir_t { DOWN= -1, NONE= 0, UP= 1 };
packet_t ptype_; // packet type (see above)
int size_; // simulated packet size
int uid_; // unique id
int error_; // error flag
int errbitcnt_; // # of corrupted bits jahn
int fecsize_;
double ts_; // timestamp: for q-delay measurement
int iface_; // receiving interface (label)
dir_t direction_; // direction: 0=none, 1=up, -1=down
x; // the node’s x location
y; // the node’s y location

….
….

}

After that you can access that x and y values from any packet that you are creating of using.

For example, before transmitting a packet you may store the x and y location of the node in the packet header structure and then send it.

If a node which receives a packet wan to know the location from which that packet originated, then it can also read it from the received packet header.

#3844 January 28, 2017 at 2:49 pm in reply to: How to use node's coordinates in AODV packets in NS-2?
"> Details
Charles Pandian
Keymaster

The average delay and average end to end delay calculation are the metrics you are mentioning in an indirect manner.
Even you can do that average calculation for any selected data flows.

Just search internet with keywords “delay”, “e2e delay” “trace analysis” awk scripts – you will find a suitable awk script for doing the measurement of your interest.

#3842 January 28, 2017 at 2:35 pm in reply to: Calculating average time taken to send a packet from source to destinatination.
"> Details
Charles Pandian
Keymaster

Check the following link. It is a similar project using same kind of routing protocol.

http://www.projectguideline.com/performance-of-swarm-routing-protocol-under-manet/

#3840 January 28, 2017 at 2:25 pm in reply to: Ant Hoc net simulation on ns2
"> Details
Charles Pandian
Keymaster

In fact, the Answer is simple, you didn’t declare the variable “ns_” in the global memory space.

you only declared a variable named “ns”.

# ==================================================== =================
# Main Program
# ======================================================================
set ns [new Simulator]

So, if you are referring an undeclared variable “ns_” from the global memory space, then you should receive the error

can’t read “ns_”: no such variable
while executing
“global $ns_”
(file “wireless.tcl” line 24)

Since there is no such variable declared in global scope.

#3838 January 28, 2017 at 2:18 pm in reply to: can't read “ns_”: no such variable – what causes this ns-2 error?
"> Details
Charles Pandian
Keymaster

For doing this, you can write a “hello Like” protocol or a beacon protocol.

 

If a mobile agent(data mules) starts to move, it should send a “hello” or “beacon” message. A sensor node which hears that “hello” or “beacon” message will start to send data to that mobile agent.

 

You can see a similar project in the following link :

http://www.projectguideline.com/a-simple-vanet-protocol-simulation-using-tcl-level-functions-of-ns-2/

In that VANET project, the stationary nodes will do something while recieving the beacon of the moving node(vehicle).

You may design your protocol based on this model.

#3837 January 28, 2017 at 2:12 pm in reply to: Data mules in ns2
"> Details
Charles Pandian
Keymaster

By default, an ns-2 packet will not contain any payload. Payload is just mentioned as size in the packet header. So, under ns-2, generally, there will not be any payload data.

So, you can not print that nothing.

 

For advanced Users :

The message passing agent of ns2 will hold little bit of sting data along with the packet – learn to use message passing agent

There exists another way to send real payload under ns-2 – but as far as I know, no one ever used that in any of the research simulation. (Some earlier video trace related projects used real payload under ns-2? – I am not sure)

#3833 January 28, 2017 at 1:22 pm in reply to: Can we view the packet content in NS2? – (not the header)
"> Details
Charles Pandian
Keymaster

In fact, labels are meant for rendering in NAM. So, generally, there is no simple way to access them from .cc code. But any way, you can do it in one way or another if you know the usage of bind command.

Bind command is used to do this tcl to cc or cc to tcl binding.

Learn about bind command.

 

 

 

 

 

#3831 January 28, 2017 at 1:12 pm in reply to: How to pass node's label from tcl to .cc in ns-2?
"> Details
Charles Pandian
Keymaster

In fact, enabling energy model in ns-2 is a simple task – nothing but a addition of few lines of tcl code. Further there is no need to set energy model for different protocols in different way.

So, if you enable energy model and run a simulation, the energy related parameters of the node will get logged along with the trace file.

So, for comparing the routing protocols and their energy consumption, you should do it in the trace analysis part – not in the tcl simulation script.

#3829 January 28, 2017 at 1:03 pm in reply to: Energy Model for DSDV and AODV routing protocols in Ad-hoc Networks under ns-2
"> Details
Charles Pandian
Keymaster

There is no ready made converter available to convert ns-2 event trace files in to Excel files.

But,  you can use space as delimit and import the ns-2 trace files (.tr files) in to Excel and then manually remove unwanted columns and rows.

But, it is not adviceable since there are LOT of freely available tcl, awk, perl, python and shell scripts available to do trace analysis in a more meaningful and logical way.

#3827 January 28, 2017 at 12:56 pm in reply to: How to import a output trace file from NS2 to excel?
"> Details
Charles Pandian
Keymaster

You should properly specify initial location, movement and size – then only NAM will render the node animation correctly.

#3825 January 28, 2017 at 12:49 pm in reply to: ns-2 Wireless Adhoc network animation not woring in NAM
"> Details
Charles Pandian
Keymaster

You can do it by modifying the .cc and .h files of the queue implementation which can be found  at

ns-allinone-x.xx\ns-x.xx\queue.

You need to add some private variables and functions to read and write those variable  in the class structure.

You should  update the value on each en-queue operation.

Instead of using IP address as index, you may use node ID.

 

 

 

#3823 January 28, 2017 at 12:43 pm in reply to: How to Count Packets Actively in Ns2
"> Details
Charles Pandian
Keymaster

In AODV under CRCN-ns-2, the channel assignment is periodically done by ChannelAssignTimer and the best channel is assigned automatically with respect to the channel condition.

The following two functions are doing this in an automatic, periodic fashion.

void

ChannelAssignTimer::expire(Event *)

{

//c a r m e n

a_->ChannelAssign();

resched(0.1);

}

 

 

//Channel  Assignment Function

void

AODV::ChannelAssign(){

//assign the interface based on greedy algorithm

//choose the channel with least interference.

//look through the route table to find the route

//with same next hop id.

….

….

}

 

So, if you need to transmit a particular packet on a particular channel, then you may need to disable this timer first and then do your way of assigning channels.

 

Charles Pandian

 

 

#3814 January 27, 2017 at 12:11 pm in reply to: How to assing a prticular channel for transmitting a particular packet?
"> Details
Charles Pandian
Keymaster

There is no ready-made or any simple way to do such clustering in MANET.

There are some sensor network extensions which contain some kind of clustering mechanism. But you can not simply use them in the way in which you want.

Further, most of such available ns-2 extensions will not address the additional overhead involved in doing such clustering under MANET or sensor network.

The main problems are message overhead and battery cost involved in the design of such clustering algorithm.

Some of the ideas that you may see in literature will use the location information of the nodes for such clustering task. But the fact is, in a practical sensor network you can not get location information without battery cost (if you use gps like functionality). Further, nodes should share the location information with one another – That will increase the message overhead tremendously. (most of the papers that you see will not even address these kind of problems in detail – it means, those algorithms will not work in a practical scenario)

 

You will not see any such complex, so called intelligent clustering based routing in most of the practically deployed sensor network.

The simple, tiny sensor nodes that were commonly used in the fields will not at all having capabilities that will be needed to code such algorithms.

The very obvious fact is, a simple, chip-level/firmware-level sleep-wakeup mechanisms available in the modern, tiny sensor network hardware itself will make the network work even better than any ingenious clustering based algorithm.

 

So, the answer to your question is somewhat complex.

There is no reliable, genuinely working and good performing clustering protocol available as a ns2 extension. (often, I realize that a well designed WSN will perform very good even without such fancy hierarchical clustering algorithm)

I saw so many students and scholars those who end-up with “FAKE” projects only because of selecting such fake/impractical ideas proposed in some papers.

Anyway, if you are really able to understand what I am telling, then you may start writing your own “genuine” clustering algorithm for ns-2 and prove that it will improve the performance of MANET or WSN and contribute something to the world.

#3812 January 27, 2017 at 11:31 am in reply to: How to do Hierarchical Clustering in MENET or Sensor Network Simulation?
"> Details
Charles Pandian
Keymaster

A background traffic is nothing but another traffic which is used simultaneously along with the primary traffic of interest.

For example, if you need to evaluate the performance of your custom voice over IP (VoIP) application, then it may use tcp or sctp  to transport your voice. But in ideal network condition, you will have all the necessary network resources and bandwidth so that you can not prove that your VoIP application will work better even in worst network condition. So, for that, you have to simulate the worst network condition with some “bottle necks” and “traffic overheads”.  For that purpose, another traffic is needed – Such background traffic may use any kind of transport agent (tcp, udp, etc.,) and may use any kind of application (cbr, vbr, telnet, ftp, etc.,)

So, for example, if you need to evaluate the performance of your VoIP application protocol with respect to different network load condition, you may do it in different ways.0

(1) Run the simulation with different background traffic conditions (against 5, 10, 20, background cbr flows) and how your fixed number of VoIP flows getting affected by such background traffic condition.

(2) Or, without any such background traffic, simply you may run the simulation with different number of VoIP flows (5, 10, 20, .. VoIP flows) and evaluate its performance

Generally, researchers prefer combining (1) and (2) to do a good evaluation while evaluating their new protocol.

#3810 January 27, 2017 at 10:56 am in reply to: What is background traffic in a network simulation?
"> Details
Charles Pandian
Keymaster

1)AODV modification:Changing the way of RREQ and RREP processing and increasing and decreasing TTL in a dynamic fasion.

There are so many probabilistic RREQ and RREP handling methods implemented  for really improving AODV similar to this idea. But for your information, the idea that you presenting will not improve the performance of AODV in any way (it will just increase the overhead in high-density network because you are increasing ttl in your idea – it will just increase the routing overhead and overall message overhead)

2.Incorporating ideas of GAF protocol by doing a better wake-up Scheduling algorithm

What you are presenting is a nothing but “same old wine in a new cup” – every wake-up scheduling algorithm will do the same in all most same way – ( I mean there is no point in your logic for any improvement better than existing  algorithms)

In a “seemingly genuine GAF implementation” that you may see in a published literature, if you see their published code, then you can find that they are simply getting the location information and grid information as follows:

 

((MobileNode *)thisnode)->getLoc(&x, &y, &z);

gid_ = God::instance()->getMyGrid(x,y);

 

It means that they didn’t address the way in which the location of nodes will be resolved in a real world scenario.

If GAF is meant for WSN, then getting location information using GPS like functionality is a also a costly thing.

Further, you can not maintain grid membership and other grid information without additional message overhead.

But in the above mentione code they are just getting them from GOD object – In real implementation there will not be any global location from which you can get those information without any additional message overhead (like using a location information server).

What I am telling is: what you “see” as a working GAF implementation is nothing but fake work.

As far as I know there is no genuine, fully working implementation available for ns2.

 

3 By dynamically changing tx powor of the nodes.

There are so many papers in the published literature that are claiming that they reduced energy by dynamically changing tx power. But most of them were fake (trust me). You will need a better understanding to realize this as fact.

Anyway I hereby try to explain it to understand its significance :

For example, if you will need 0.01 w of energy to transmit a packet up to 100 meters, and you will need another 0.01 w to relay the packet to another 100 meters.

But if you try to transmit the packet for 200 meters, then you can not do it with 0.02 w. You will need higher than 0.02 watt. It is the nature of radio waves.

For example, if we need to transmit a packet up to 200 meters, then it may need (for example) 0.03 w. But if you do it in two steps (I mean two hops) with 0.01 watt, then totally you will need only 0.02 w. It is actually true.

But “reducing tx power technique” under high-density MANET or WSN will never practically work. Because, if you increase the hops, then it will increase the number of broadcasts. The increase in broadcasts will lead to more collisions and re-transmissions – It will lead to a high end to end delay and very poor throughput. So, there will be a lot of RREQs and RREPs in the network and will actually consume much power in discrete steps to find route in the large multihop network.

Since there will be a considerable increase in both transmissions, as well as receptions  (because in wireless network, if one transmits, all the neighbors will receive it (yes receiving of too much packets also drain the batteries)

So, as per your imaginations, you can not reduce the overall energy consumption of network by reducing or increasing power.

So simply the things that you are telling in (3) will not improve performance in any way or reduce the overall energy consumption of the network (you may need to do lot of genuine experiments to really understand what I am trying to say).

You may see a lot of such idiotic ideas in reputed journal papers. Simply, they are all nothing but manipulated,  fake works (or only work with some funny network constraints – and simply will not give improvement under real networking condition).

 

Charles Pandian

 

#3640 January 21, 2017 at 1:23 am in reply to: Changing the way of RREQ and RREP processing, dynamically changing txpower of the nodes under ns-2
"> Details
Charles Pandian
Keymaster

What is the best way you think to update the node position and number of neighbours? I need an effective way from your experience and knowledge.

The hello mechanism is the only effective way to get update about a node’s current position(if we are not considering a centralized location service).
Particularly, in AODV, there is an existing hello mechanism – you just need to activate it.

What is the best way to estimate the number of neighbours in the range to use enough power and also to reach them? I need an effective way from your experience and knowledge.

Just by receiving the hello replies from the neighbors is the only effective way to estimate the number of neighbors.
Again, in AODV, there is an existing hello mechanism – so that it will not increase the overhead while implementing neighbor discovery.
Over hearing the transmissions of neighbors at application layer is also the another way to count neighbors – but for that you may need to bring a node to promiscuous mode which will increase the overhead in all layers. I mean, you need to process all the packet at mac layer and send that information to Application layer.

What is the best way to calculate how many neighbours each node has? How can you calculate the optimum number of neigbors? I need an effective way from your experience and knowledge.

As far as MANET is concerned, there is no reliable way to calculate how many neighbors a node has – but over hearing the transmissions of neighbors is the another way to count neighbors.

What is the best way to estimate the distance and calculate that between nodes? I need an effective way from your experience and knowledge.

If you really know the absolute location of  two nodes, then the simple Euclidean distance calculation is enough. But there is no reliable way to calculate the absolute location of a node at any instant under a mobile condition.

What is the best way to adjust the transmission power? How can you select the optimum transmission power? I need an effective way from your experience and knowledge.

Changing TxPower in wireless Physical is the simplest way to change transmission power.
There is no reliable way for deciding  “optimum transmission power” – it will depend upon several factors – even those several factors also dynamically change in a highly mobile, high density network.
(that is why most of the dynamic tx power based AODV algorithms that you see in published paper will not really give any improved result in you try to implement it)

The only thing I know for reducing overall energy consumption is – reducing overhead in route request and reply mechanisms and  using more reliable paths to avoid link failures. Because, routing overhead is just using energy – link failures also initiate re-route discovery – so that consume energy again and again just for route discovery
(Here I am not mentioning the power used for transmitting actual data payload).

There are some funny algorithms you may see in the some published works based on reducing tx power with respect to local node density – trust me – they will not really work as it seems – Because, just decreasing the tx power with respect to the neighbor count will again reduce the neighbor count in the very next instant – so, it will just cause link failures at the very next instant of time and re-initiate route discovery and will increase the routing overhead and message overhead  and all kinds of other overhead to a very higher level.

So the following is my simple idea related with any genuine modified AODV implementation- I hope this may work

1. Keep tx power constant (a fixed value) for all  hello requests and replies- and  all route request and reply messages.
2. control the forwarding of  route request and reply messages based on the neighbor count at that node – this will help to establish more stable /reliable paths (and will reduce little bit of energy).
2. While routing a data packet through a neighbor, just change the transmission power to a minimum level to reach only up to that node (I mean, a assumed good, minimum distant) – after transmission of data packets, change the tx pwer to the original fixed value (for transmitting other kinds of packets)  so that normal working of routing will not get disturbed by any reduction in tx power. (this will also reduce little bit of over all energy consumption of the network).

Charles Pandian.

#3638 January 21, 2017 at 1:15 am in reply to: Getting Node Position, Neighbors, Distances and Transmission Power etc., Under ns-2
"> Details
Charles Pandian
Keymaster

Dear Scholar,

It depends upon the exact research problem that you are working on.

The reason is:
Flooding can be implemented on application layer or network Layer or any other layer depending upon the need.

For example, there is a flooding based route discovery mechanism already implemented in AODV.  It is often proved that the use of flooding in routing protocols creates a lot of overhead in a high-density MANET. So, if your research is about reducing such overhead in a network routing protocol by improving flooding or replacing flooding at routing layer, then you may implement your ideas on network layer. For that you may simply try to reduce the overhead of flooding at routing layer by designing an intelligent, controlled flooding algorithm or entirely replace flooding based route discovery with any other similar and efficient algorithm – In this case, you have to implement it as a C++ code by modifying aodv.cc and aodv.h (and may be few more files)

On the other hand, if you are just doing a research only on a different family of message flooding like algorithm, then you can implement it as a stand-alone algorithm in application layer itself – In this case, you may implement it in the main simulation tcl code itself or as a separate C++ level application agent( .cc and .h files).

#3634 January 21, 2017 at 12:40 am in reply to: Implementation of flooding and other related protocols on ns-2
"> Details
Charles Pandian
Keymaster

Dear Scholar,

For a simple discussion, you may ask questions in this forum after getting free registration.

But, if you need exclusive personal assistance, then you may communicate through e-mail.
For your information, I am providing exclusive personal assistance for scholars those who already selected a research problem and got registered for their Ph.D in some University.
So, if you already registered a research topic and finding trouble in identifying a specific problem, then of course, I can assist you.

#3632 January 21, 2017 at 12:02 am in reply to: Can you guide me to identify a problem in this research area?
"> Details
Charles Pandian
Keymaster

Pegasis implementation is available  at the following link

https://drive.google.com/file/d/0B7S255p3kFXNUDB4VmI4ZVdZTjg/view?usp=sharing

after installing it you may  use it to run the Pegasis simulations. 

#3016 December 19, 2016 at 8:19 pm in reply to: PEGASIS Protocol implementation in NS2
"> Details
Charles Pandian
Keymaster

In ns-2, there are some sample code available for simulating simple flooding in wired as well as wireless domain. One can customize any one of such algorithm and design a time synchronization algorithm. Even the route discovery mechanism available in routing protocols use flooding. So, it is also possible to implement a time synchronization method on suitable routing algorithm.

#3013 December 19, 2016 at 7:18 pm in reply to: time synchronization code…
"> Details
Charles Pandian
Keymaster

Generally you can not get any such readily usable code for such complex implementation from anywhere in the  world.   But you can develop such code from scratch.

Further, as far as I know, there is no Lossless Hiding in Images. Even in all kind of Steganographic techniques and Watermarking techniques, the host image will get distorted to some minimum level(at least).

Stepts to Achieve your Goal :

1. First of all try to simplify your goal as follows – I mean remove the complex things from it.

  • Lossless and Reversible Data Hiding in Encrypted Images with Public Key Cryptography
  • Lossless and Reversible Data Hiding in Encrypted Images with Public Key Cryptography
  • Lossless and Reversible Data Hiding in Encrypted Images with Public Key Cryptography
  • Lossless and Reversible Data Hiding in Encrypted Images with Public Key Cryptography

2. Finally you end up with a simple thing :

           Data Hiding in Images

3. You may find many freely available Data Hiding algorithm/code.

4. Start your project with a simple and suitable  code for “Data Hiding in Images.”

5 After successfully implementing a simple data hiding model, you may try to extend it.

6. Now you may improve it by adding “retrieval part” and do a  “Reversible Data Hiding in Images”

7. And then you may improve it as “Reversible Data Hiding in Images with Public Key Cryptography”

8. But I doubts that there is no way to do a “Lossless and Reversible Data Hiding in Encrypted Images with Public Key Cryptography”

 

Simple way to develop such Image processing based application is prototyping it with Matlab software.

Search Internet with keyword “steganography code in matlab” or “steganography code in C++” – you will find lot of such code – you may start your project by selecting one such simple and suitable code.

 

#2991 December 17, 2016 at 5:06 pm in reply to: Data Hiding in Images
"> Details
Charles Pandian
Keymaster

ok. There is no simple answer for your one line question.

Just start learning ns2 by following the steps in the following post

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

#954 April 18, 2014 at 1:26 pm in reply to: How to write ns2 program for this topic
"> Details
Charles Pandian
Keymaster

Set the required error rate in a variable

set ErrorRate 0.15

write a procedure to set error rate

proc UniformErr {} {
global ErrorRate
set err [new ErrorModel]
$err unit packet
$err set rate_ $ErrorRate
$err drop-target [new Agent/Null]
return $err
}

Call the procedure while creating the MANET node.

ns node-config -adhocRouting $val(rp)
-llType $val(ll)
-macType $val(mac)
-ifqType $val(ifq)
-ifqLen $val(ifqlen)
-antType $val(ant)
-propType $val(prop)
-phyType $val(netif)
-topoInstance $topo
-agentTrace ON
-routerTrace ON
-macTrace ON
-movementTrace ON
-IncomingErrProc UniformErr

#955 April 18, 2014 at 1:24 pm in reply to: how to introduce error model in manet?
"> Details
Charles Pandian
Keymaster

Generally, you can do a minimal install of Cygwin (I mean some selected components of Cygwin) under Windows 7 (or higher). But, installing a fully functional ns2 will not be possible with that Windows 7 install of Cygwin. (there seems some OS architectural limitation which will raise some conflicts)

But, you can install Windows XP under Virtualbox or any other visualization environment under Windows 7 and then install Cygwin and ns2 under that virtual Windows XP installation.

It will run without any problem and of course, after doing some fine tuning (on XP), you may get near native hardware performance under virtual Windows XP.

Even you may install any linux variant on the Virtualbox under Windows 7 and install ns2 on it – it will work even better than previously mentioned WinXP+Cygwin+ns2 install.

#828 January 20, 2014 at 4:28 am in reply to: to install ns2 2.34
"> Details
Charles Pandian
Keymaster

What is the output of the following patch operation?



patch -p1 < dymoum/dymoum_ns-2.34_v0.3.patch

For example, if you receive any single error like this

Hunk #1 FAILED at XX

then it means that the patch operation has failed.

If you compile ns2 after a partial patch operation or a failed patch operation, then you can not compile ns2 successfully again.

#952 July 18, 2013 at 2:35 pm in reply to: how to patch DYMOUM in ns2.34
"> Details
Charles Pandian
Keymaster

ns2 has some sensor network routing protocols under the folder
ns-allinone-2.xxns-2.xxdiffusion
you may use them.

Or, you can design your own sensor network by using the features available for simulating wireless networks/MANET.

Or you may use a suitable extension such as Mannasim
(http://www.mannasim.dcc.ufmg.br/)

#949 April 24, 2013 at 7:51 pm in reply to: Hoow to design and simulate Wireless sensor network in ns2 ?
"> Details
Charles Pandian
Keymaster

In ns2, there is an agent called MessagePassing Agent. You can pass real data (in the form of characters) using this agent. The default message size will be 1500 bytes
(defined in ns-default.tcl as
Agent/MessagePassing set packetSize_ 1500).
But you may increase it as per the requirement.

But, I hope that you can not use it over TCP.

#951 April 24, 2013 at 7:43 pm in reply to: How to add data payload
"> Details
Charles Pandian
Keymaster

The resent versions of ns2 only contains multi-path AODV (AOMDV). But, MAODV is multicast AODV.
MAODV is available as an extension for some old version of ns2.

#943 December 9, 2012 at 5:22 pm in reply to: maodv in ns 2.34
"> Details
Charles Pandian
Keymaster

The default ns2 does not have any data aggregation mechanism. But, the sensor network extension called mannasim is capable of doing data aggregation.

#941 December 21, 2011 at 6:46 pm in reply to: data aggregation in ns2
"> Details
Charles Pandian
Keymaster

Open any text editor and key-in your ns2/tcl simulation script and save it with extention “tcl” like experiment1.tcl.

After that, you can run it from the terminal window ( here we assumed that ns2 is already installed in that machine)

# ns experiment1.tcl

Writing C++ level agent or modifying an existing agent/protocol and compiling ns2 is little bit difficult than that of writing and running a tcl level simulation script . So, you can not get any simple “step by step” answer for your second question. Learn it from appropriate tutorials.

You may start learning it from the following link
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

#939 September 9, 2011 at 11:03 am in reply to: NS2 PROBLEM: Where to type codes?
"> Details
Charles Pandian
Keymaster

There is no ready-made / simple way to do clustering in MANET.
There are some sensor network extensions which contain some kind of clustering mechanism. But you can not simply use them the way in which you want. Further, as far as I know, in ns2, security aspects of the simulations were not considered in most of the simulations. The fact is, (according to my belief) implementing a real encryption or decryption in a simulator is beyond the scope of the simulator. Even security aspects also should be simulated in one way or another.

But, Qualnet seems to to be having some security aspects (like IPsec, RSVP).

#934 February 25, 2011 at 8:04 am in reply to: Clustering in MANET
"> Details
Charles Pandian
Keymaster

It is very simple.

set Locationx [$node(1) set X_ ]
set Locationy [$node(1) set Y_ ]
set Locationz [$node(1) set Z_ ]

#935 February 25, 2011 at 7:55 am in reply to: Getting Coordinates of a node in ns2 using tcl
"> Details
Charles Pandian
Keymaster

I hope that it is possible.

Before creating the nodes, we generally supply all the common node parameters.
Instead of that, you may create nodes one by one with different node parameters.

#937 February 25, 2011 at 7:51 am in reply to: can we change the buffer size for each and every nodes in ns
"> Details
Charles Pandian
Keymaster

What do you mean by”mesh for microprocessor”?

#932 January 31, 2011 at 1:49 pm in reply to: Mesh
"> Details
Charles Pandian
Keymaster

In fact you can call/use any external program during a ns2 simulation. So even you can call the compiled version of your c/c++ code from tcl scripts.


But, it is not the generic/correct way of writing a protocol/agent for a ns2 related network simulation.

If you want to write your c/c++ code, then you may incorporate them inside any of the C++ protocol/agent code of ns2. Or write them separately and create your own protocol/agent for ns2. For that you have to follow all the standards of writing a ns2 C++ code.

In other words, there can be only one “main()” function in ns2 code structure. So you can not have another one inside your C++ code.

Sorry for confusing you. – In fact, it may take some time to explore and feel the beauty of ns2 and its C++-Tcl duality nature, and of course, may take years to fully understand and practice/use this.

#931 January 24, 2011 at 7:39 am in reply to: linking c++ in ns2
"> Details
Charles Pandian
Keymaster

There is a patch available for simulating Wireless Mesh Networks in ns-2.

http://cng1.iet.unipi.it/archive/ns2mesh80216/ns2mesh80216-2.33-081113.patch.gz
http://cng1.iet.unipi.it/archive/ns2mesh80216/ns2mesh80216-2.31-071030.patch.gz

Further, it seems that there is some example script available with that patch for simulating mesh network.

Read the following link for further information

http://cng1.iet.unipi.it/wiki/index.php/Ns2mesh80216

#930 December 6, 2010 at 9:04 am in reply to: simulation of WMN in ns2
"> Details
Charles Pandian
Keymaster

Use -fmtx option to specify the required format.

#928 October 28, 2010 at 11:32 am in reply to: xgraph in NS2
"> Details
Charles Pandian
Keymaster

I hope that num_connect is not a default parameter related with god object. So instead of using “$god_ set num_connect” , just use the total number of nodes or any number representing node count (like $val(nn) )in that calculation

#927 October 28, 2010 at 4:15 am in reply to: Problem with God object!
"> Details
Charles Pandian
Keymaster

Read this post.

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=386

#926 October 17, 2010 at 4:44 am in reply to: NS2 interfacing with matlab
"> Details
Charles Pandian
Keymaster

Yes. It is possible; but not a simple thing.
Read the procedure available in this link
http://www.cse.msu.edu/~wangbo1/ns2/nshowto3.html

#925 October 7, 2010 at 4:36 pm in reply to: Calling Matlab function in ns2!!!
"> Details
Charles Pandian
Keymaster

Hopefully, there will not be a easy solution for your problem.

But, “A Linux TCP implementation for NS2” is providing easy way to integrate linux implementations of tcp along with ns2.
http://netlab.caltech.edu/projects/ns2tcplinux/ns2linux/index.html

Even you can easily add your own algorithm in a simple way.

Higher versions of ns2 contains tcplinux by default.

You may try that also.

#924 October 7, 2010 at 5:13 am in reply to: Help For Quick Start TCP Implementation in ns 2.31 / ns 2.34
"> Details
Charles Pandian
Keymaster

I think that there is no such protocol readily available for ns2.

class LL/Sat is currently the only one available in ns2. Refer Chapter 17 of ns2 manual for more information.

You have to extend/modify any available Link-Layer Protocol of ns2 and with AX.25 Amateur Packet-Radio Link-Layer Protocol specifications
http://www.tapr.org/pub_ax25.html

But it is not a simple task – It may take months/year or more to complete it successfully.

#920 October 7, 2010 at 3:08 am in reply to: AX.25 simulation on NS-2
"> Details
Charles Pandian
Keymaster

I think, you have to use a patch like
http://www.matlab.nitech.ac.jp/~khpoo/research/index-poisson.htm
or
http://www.iis.sinica.edu.tw/~cclljj/misc/notes/poisson.cc.html

and then customize it according to your need.

#923 October 7, 2010 at 2:57 am in reply to: Looking for a poission flow generator in ns2!!
"> Details
Charles Pandian
Keymaster

The correct syntax of “$ns create-trace Drop file s0 d0” would be like the following one :

$ns create-trace Drop $outfile $s0 $d0

Where
$file is the file descriptor ( which we initiated like this : set outfile [open out.dat w]

$s0 is the source node
$d0 is the destination node

#922 October 7, 2010 at 2:49 am in reply to: how to use $ns create-trace type file src dst op
"> Details
Charles Pandian
Keymaster

Read similar discussions in the following posts.

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&p=101
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=346

some of the simple errors of this kind can be resolved by modifying the code little bit (like using type casting). Otherwise we have to change the gcc version if we face LOT of such errors during compilation.

#921 October 6, 2010 at 6:16 pm in reply to: Problem with installation of Ns 2.29 under Fedora
"> Details
Charles Pandian
Keymaster

Before starting to study about VANET routing protocols, first try to simulate normal routing protocols under VANET mobility models.

You can start with Manhattan Grid Mobility model.

#916 October 6, 2010 at 5:40 pm in reply to: Help: vanet routing algorithms
"> Details
Charles Pandian
Keymaster

Try to install lower version of gcc (and make) and then try compiling ns2.
It may solve your problem.

#915 September 24, 2010 at 8:50 am in reply to: error in running ns in fedora 12
"> Details
Charles Pandian
Keymaster

start from the following link
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

#914 August 24, 2010 at 1:32 pm in reply to: Beginner
"> Details
Charles Pandian
Keymaster

Generally, any agent of ns2 will receive a packet through a function named ::recv. For example the following aodv function will receive packets at routing layer of aodv.



void

AODV::recv(Packet *p, Handler*) {

.....

...

After receiving the packet it will behave with respect to the received packet.
It may forward or send or drop the packet at that point (in side that function body).

On the other hand, timers also can be called periodically to do some routine function. So at at point also sending a packet can happen.

So just search for lines like



Scheduler::instance().schedule(target_, p,

0.01 * Random::uniform());

Actually, the function “schedule” will schedule a packet for transmission.

Technically, what I am saying is, you can not count the transmitted packet from one point.

[/b]

#911 August 4, 2010 at 11:45 am in reply to: How can i calculate the no of data packets sent out from my
"> Details
Charles Pandian
Keymaster

The error “SplitObject unknown” signifies that you not yet completed the patch correctly and compiled ns2 successfully with swan.

#912 August 4, 2010 at 11:34 am in reply to: error while running swan
"> Details
Charles Pandian
Keymaster

open menu.lst using a text editor (like vi) and revert the changes.
Probably it will work. (boot the system with a bootable live linux CD and mount the appropriate device which contains the modified version of menu.lst and do the editing)

#913 August 4, 2010 at 11:31 am in reply to: mipv6 Installation problem in linux (ubutnu)
"> Details
Charles Pandian
Keymaster

According to the 802.11e documentation,
changes to 802.11e parameters (CW_MIN, CW_MAX, AIFS, TXOPLimit, PF) should be made in /ns-allinone-2.xx/ns-2.xx/mac/802_11e/priority.tcl for each queue. After changing this file then make ns2 again. Then only the parameters will change.

#904 July 17, 2010 at 4:24 am in reply to: How to change parameters via TCL file in 802.11e TKN model
"> Details
Charles Pandian
Keymaster

That is the very basic way of simulating DoS attack in a ns2 simulation.

If you run the simulation, it may create a trace file. For example, the following lines will open the file TCP_Name-out.tr and write all trace events in it.



set TraceFD [open $TCP_Name-out.tr w]

$ns trace-all $TraceFD

After running the simulation, you may use a suitable scripting language like awk, perl etc., etc., to analyze the trace file. Or simply you may use shell scripts to do trace analysis. Search Internet you may find lot of such awk and perl files for trace analyzing the ns2 trace file.

You may learn about trace analysis from the following tutorials.

1. The Famous ns2 Tutorial by Marc Greis http://www.isi.edu/nsnam/ns/tutorial/
2. NS by Example by Jae Chung and Mark Claypool http://nile.wpi.edu/NS/

#907 July 17, 2010 at 4:16 am in reply to: DOS and DDoS Simulation +Comparison and Analysis of Xgraph
"> Details
Charles Pandian
Keymaster

How did you simulate DDoS in ns2?

#905 July 14, 2010 at 10:11 am in reply to: DOS and DDoS Simulation +Comparison and Analysis of Xgraph
"> Details
Charles Pandian
Keymaster

certainly it will not be a problem with ns2 or C++ . your logical error will be the cause of this error (logical error).


Actually, the question was not clear.


“i notice that if my printf are within if else statements which are in a function(testingprint()), only node 0 will be printed
but if i remove the if else statements, all my nodes will print.”

You are checking the value of ff2 in the if statement. so in the first case, ff2 seems to be having value 0 if and only if the index=0. So it is printing for node 0 only.

If you are removing the if, then all it will not consider the value ff2 and so may print everything.
[/i]

#902 June 23, 2010 at 3:56 am in reply to: Strange problem with ns2 printf
"> Details
Charles Pandian
Keymaster


Agent/Ping instproc recv {from rtt} {

$self instvar node_

puts "node [$node_ id] received ping answer from $from with round-trip-time $rtt ms."

}

If the PningAgnet is receiving a ping reply, then the control will enter this block of code and this print command(puts) will print that info (in console/terminal). “from and rtt” are the two parameters supplied to this function.

#901 June 16, 2010 at 5:15 am in reply to: tcl code
"> Details
Charles Pandian
Keymaster

If you are able to compile the same thing successfully while login as another user, and can not do the same in a particular user account, then certainly it will be related with user right on a particular directory/command and some environment settings like the PATH environment variable. Just explain each step of your installation/compilation.

#891 June 14, 2010 at 3:07 am in reply to: problem with ns2.34 installation
"> Details
Charles Pandian
Keymaster

try after removing the following lines or put ‘a #’ in front of these lines . (particularly I am having doubt in the first line – it seems to be changing the cpp environment – so the default gcc will be ignored)



export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/home/avinashc/Download/LEDA/incl

export LIBRARY_PATH=$LIBRARY_PATH:/home/avinashc/Download/LEDA

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/avinashc/Download/LEDA

export TOSROOT=/opt/tinyos-2.x/

export CLASSPATH=$TOSROOT/support/sdk/java/tinyos.jar:.

#export MAKERULES=/opt/tinyos-2.x/support/make/Makerules

export TOSDIR=/opt/tinyos-2.x/tos

export PYTHONPATH=$PYTHONPATH:$TOSROOT/support/sdk/python

After that add appropriate lines for ns2.
(Read about setting values in profile using the following link.
http://www.projectguideline.com/modules.php?name=News&file=article&sid=9)

#889 June 13, 2010 at 3:00 pm in reply to: problem with ns2.34 installation
"> Details
Charles Pandian
Keymaster

Just remove the following in that if statement.
&& $35 ==”message”

Because I hope that it will not be necessary. Because all the packet sent or forwarded at the network layer should be taken in to account. So no need to add any other filter at that point.

Charles Pandian.

#897 June 12, 2010 at 6:08 pm in reply to: DSDV in new trace format ??
"> Details
Charles Pandian
Keymaster

This may be due to the wrong path/environment setting in the profile of that particular user (.bash_profile – which is a hidden file – which sets some environment variables -ns2 installation will add few lines in to it – or you may be set some thing wring in that profile )

#ls -a will show all the files including the hidden ones.

use vi or any other editor to edit it.

#887 June 12, 2010 at 6:03 pm in reply to: problem with ns2.34 installation
"> Details
Charles Pandian
Keymaster

Read similar discussions in the following posts.

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&p=101
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=346

some of the simple errors of this kind can be resolved by modifying the code little bit (like using type casting). Otherwise we have to change the gcc version if we face LOT of such errors during compilation.

#885 June 11, 2010 at 11:57 am in reply to: problem with ns2.34 installation
"> Details
Charles Pandian
Keymaster

General, if you will try to install a old version of ns2 on a new version of linux/unix, then you will face lot of this kind of errors. This is mainly due the the default(higher) version of gcc (compiler) present in that linux/unix environment.

To avoid this kind of errors, simply try to use old version of gcc for compiling old version of ns2. (learn to remove the default version of gcc and learn to install a old version of gcc )

In the following link, you may find a similar problem

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&p=101

#894 June 11, 2010 at 11:50 am in reply to: Error While Installing ns-allinone-2.1b6
"> Details
Charles Pandian
Keymaster

Learn about energy model of ns2 from the Official ns2 Manual – http://www.isi.edu/nsnam/ns/doc/index.html
http://www.isi.edu/nsnam/ns/doc/node225.html

Learn about initializing energy model in a tcl simulation script. (it should be done before the creation of the node.

set the default node configuration as follows (refer other scripts from internet)



$ns node-config -adhocRouting $val(rp) 

-llType $val(ll) 

-macType $val(mac) 

-ifqType $val(ifq) 

-ifqLen $val(ifqlen) 

-antType $val(ant) 

-propType $val(prop) 

-phyType $val(netif) 

-channelType $val(chan) 

-topoInstance $topo 

-agentTrace ON 

-routerTrace ON 

-macTrace ON 

-movementTrace ON 

-energyModel $val(engmodel) 

-initialEnergy $val(initeng) 

-txPower  $val(txPower) 

-rxPower  $val(rxPower) 

-idlePower  $val(idlePower)

Learn to extract the power of a particular node or all the nodes which was logged in event trace file (learn trace analysis and learn to use awk/perl scripts for trace analysis)

😉 Yes. It is not a simple thing 😀

#895 June 11, 2010 at 11:36 am in reply to: Calculating Energy efficiency in wsn (smac)by Sleep & Aw
"> Details
Charles Pandian
Keymaster

There is a patch available for tora. (search this forum – I mentioned in some other post)

You can make it work without any error after the patch operation.

But, as far as I have experimented with tora, it will not produce results which were mentioned in earlier papers (some of those comparison graphs in some previous papers) . Or even it will not able to send a single packet in a network of about 10 mobile nodes. If you run a 10 node scenario randomly for 10 or 20 times, the delivery of packet will occur very very rarely.

Even I tried it several times with different older versions of ns2 – but the results were same. (old versions of ns2 will not require any patching for tora)

Even it you are running the example available in ns2, ( a four/few node scenario), the transmission will happen only in a particular orientation of the nodes. The transmission will not happen in some other orientation even if the nodes were in communicable range.

Is it a bug in default implementation of tora? or is it due to my ignorance?
I do not know.

AVOID SELECTING TORA FOR FURTHER ENHANCEMENTS (as a research area)

#893 June 11, 2010 at 11:23 am in reply to: problem with tora after patch up
"> Details
Charles Pandian
Keymaster

The download link seems to be a broken one. Contact the Author for another link to download it.

or you may download(copy-past) separate files from the following google code search link.(but it will need lot of patience and experience to make it work)

http://www.google.com/codesearch/p?hl=en#klyNBxxjEYE/software/owns-src.tar.gz%7CnnY0IGacGcg/OWNS/ns-2/cbr_session_traffic.cc&q=owns-src&d=2

#882 May 26, 2010 at 2:54 pm in reply to: IP over DWDM Simulation using NS2 to Measure The Performance
"> Details
Charles Pandian
Keymaster

This info is from a tcl simulation code. You may thank the original (unknown) author for this information.



#Pt_		Distance



#0.2819		250m

#0.5845		300m

#1.0828		350m

#1.8472		400m

#2.959		450m

#4.5099		500m

#These Pt_ values correspond to the default RXThresh_ value (3.652e-10).

#They are not super-accurate, so the transmission range may differ by +/- a meter or two.
#879 May 25, 2010 at 7:04 pm in reply to: the formula to convert from w to meter
"> Details
Charles Pandian
Keymaster

There is a WDM extension available for ns2.
http://www.eecs.wsu.edu/~dawn/software/owns.html

But I hope that DWDM can not be simulated on ns2 or Matlab since it seems there is only limited support for this in both of this platforms.

There seems a possibility to simulate DWDM networks with OPNET simulator (with SP Guru® Transport Planner).
But I am not sure abouts it default availability with the freely available student version of the OPNET simulator. (kindly post a reply if I am wrong)

http://www.opnet.com/solutions/network_planning_operations/spguru_transport_planner.html

#880 May 25, 2010 at 7:00 pm in reply to: IP over DWDM Simulation using NS2 to Measure The Performance
"> Details
Charles Pandian
Keymaster

I think, for this you need not create another node with same id. Just alter the packet header and transmit it (fill it with whatever the source and destination id you want). But, I think, it can only be done from a c++ level agent code in ns2 (not from tcl level code).

#877 May 15, 2010 at 3:57 am in reply to: how to do cloning in ns2 for wsn?
"> Details
Charles Pandian
Keymaster

Briefly explain the communication scenario. It may help to find an alternate way of doing it.

#875 May 7, 2010 at 4:52 pm in reply to: how to do cloning in ns2 for wsn?
"> Details
Charles Pandian
Keymaster

I hope that, in ns2, it is impossible to create more than one node with same ID.
So try your simulation in some other indirect way.

#873 May 7, 2010 at 3:33 pm in reply to: how to do cloning in ns2 for wsn?
"> Details
Charles Pandian
Keymaster

This is a common error received during compiling (making) ns2 after some modification.

It means you are doing some copy-past operation in makefile. The copy-past operation will add some non standard end of line characters at end of those lines. The make command will fail if it will ‘see’ those strange end of line characters. So, you can remove those unwanted characters using a suitable text editor(lean it).

Or simply type those new lines in makefile using a simple editor like “vi”(lean it).

It will solve the problem.

#866 May 7, 2010 at 3:25 pm in reply to: Wimax package with ns_2
"> Details
Charles Pandian
Keymaster

This error(SplitObject unknown) tells that you not yet successfully compiled ns2 with your new protocol/agent. So, if you are referring your new protocol/agent in a tcl simulation code, it will say “SplitObject unknown” – in your case, any reference to any of the procedure/function of your protocol will raise this error.

It means, you not yet successfully integrated your protocol with the default ns2.

#878 May 7, 2010 at 3:18 pm in reply to: Error while running my own protocol
"> Details
Charles Pandian
Keymaster

you did not show the body of the procedure “plotWindow”. (So it can do anything)

Generally, the xgraph command will expect two column input (from a file or terminal). The first column is ‘x’ and the second column will be ‘y’. So the unit of x and y axis will depend upon the thing you are having in ‘x’ and ‘y’ columns.

#872 May 7, 2010 at 3:13 pm in reply to: what this procedure role in tcl files
"> Details
Charles Pandian
Keymaster

the command “$ns_ node-config” can be used to set the parameters of a node during node creation. (-txPower , -rxPower will decide the transmission range and coverage)



$ns_ node-config -adhocRouting $opt(adhocRouting) 

-llType $opt(ll) 

-macType $opt(mac) 

-ifqType $opt(ifq) 

-ifqLen $opt(ifqlen) 

-antType $opt(ant) 

-propType $opt(prop) 

-phyType $opt(netif) 

-channelType $opt(chan) 

-topoInstance $topo 

-wiredRouting OFF 

-agentTrace ON 

-routerTrace ON 

-macTrace ON 

-energyModel $val(engmodel) 

-initialEnergy $val(initeng) 

-txPower  $val(txPower) 

-rxPower  $val(rxPower) 

-idlePower  $val(idlePower)

if you create a node after this configuration, then it will create the node with that default parameter.



set node_($i) [$ns_ node]

After creating a node, if you want to create another node or another set of nodes with different parameters, then again use the same command “$ns_ node-config” with different parameters and then use “set node_($i) [$ns_ node]”
($i will differ and denote a node)

#815 May 7, 2010 at 3:08 pm in reply to: Transmission Range
"> Details
Charles Pandian
Keymaster

I hope that, AOMDV – a multi-path routing agent is available for ns2. But SMR is not available from any contributed code of ns2.

#871 May 7, 2010 at 2:59 pm in reply to: Need code for SMR urgent
"> Details
Charles Pandian
Keymaster

there is nothing special with file names (out.tr,win.tr etc.).

But to learn about trace analysis, you may start with the following links

http://nsnam.isi.edu/nsnam/index.php/NS-2_Trace_Formats
http://bunlab.twbbs.org/ns2_trace_format.htm
http://www.winlab.rutgers.edu/~zhibinwu/html/ns_trace.html
http://mailman.isi.edu/pipermail/ns-users/2003-October/036546.html (but now, it seems that tracegraph is not available for download)

Further, there are so many awk, perl and other scripts freely available for trace analysis.

#870 May 7, 2010 at 2:56 pm in reply to: how to read trace file format in ns-2
"> Details
Charles Pandian
Keymaster

If you mean to print something in console, then you may use



fprintf(stderr, "n%4.3d", x);
#869 May 7, 2010 at 2:50 pm in reply to: Which function to get it a number with some digits
"> Details
Charles Pandian
Keymaster

Read this post and learn to recompile ns2.

http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&p=157

#868 April 26, 2010 at 8:26 pm in reply to: how to compile new protocol in ns-2
"> Details
Charles Pandian
Keymaster

Just use



fprintf(stderr,"your text string");

or

to display an integer along with a text, use



fprintf(stderr,"The number is : %d", 10 );
#867 April 25, 2010 at 7:52 pm in reply to: to show a string or number in terminal write by C++ in NS2
"> Details
Charles Pandian
Keymaster

While compiling ns2, finally, it(the make command) will generate a binary executable file in the name “ns”. If you run that command you will have % prompt. (just similar to a tcl shell prompt).

Ok.

ns is made up of thousands of such individual C++ files (*.cc)

In C++ (and in c) there is a convention that the some constants, variables, function prototypes and the class objects were declared in a header file(*.h) with the same name of the c++ file which will use those constants, variables, functions and classes.

While compiling(making) ns2 using make command, all those refereed C++ files will get compiled individually and the corresponding, generated object files (binary) will be finally linked (using the linker) which will produce the final binary file “ns” (in windows, it will be ns.exe)

Note :
You can not use them (C++/h) separately just like a normal C program which you will write in your C++ practicals class/lab. You can not run them individually since there will not be “main()” function in those separate code segments.

#865 April 15, 2010 at 8:03 pm in reply to: new user
"> Details
Charles Pandian
Keymaster

I can not understand what your are exactly asking.

But in Chapter 7 “Queue Management and Packet Scheduling” , section 7.3, the ns manual defines it as follows:

bdepartures_ Running total of bytes contained in packets that have departed (not dropped).

Refer that chapter for more information.

#862 April 10, 2010 at 7:02 pm in reply to: bandwidth analysis
"> Details
Charles Pandian
Keymaster

What is multi-part OLSR?

#861 April 6, 2010 at 7:32 pm in reply to: Could anybody has code of multi-part OLSR in ns2?
"> Details
Charles Pandian
Keymaster

I thing you are just trying to download ns2 from sourceforge.net (you are actually not installing it from the browser).
Just to solve this download problem, use a better browser like Firefox or Opera.

Further, even after successful download, you can not install ns2 directly on windows. You will need cygwin environment (a unix like environment for windows).

Most of the (Old) versions of cygwin can not be installed on windows vista and windows 7.

So install cygwin on windows xp only.

There may be driver problems while installing windows xp on latest laptops ( and desktops too)

#860 April 4, 2010 at 1:37 pm in reply to: Error while installing NS2.34
"> Details
Charles Pandian
Keymaster

Probably, no one will provide you any ready-made answer/code to this question.

According to my understanding, this can be implemented at tcl level ns2 simulation itself. For that, you can extend the class Agent/Messagepassing to behave as your protocol. (you have to extend the send and recv functions in your tcl code and write the protocol).

If you want to do it in a hard way, then you can do the same thing in c++ level code (you can extend the fuctionalities of messagepassing.cc and messagepassing.h)

Or you can write a new application level Agent or routing agent to handle this kind of communication scenario – but it will be more complex than the above two.

#855 April 2, 2010 at 5:52 pm in reply to: Whitelist implementation in ns2
"> Details
Charles Pandian
Keymaster

Sorry.

The possibilities of these kind of agent/object related errors will need lot of time to figure out.

#859 April 2, 2010 at 5:42 pm in reply to: list of tcl object
"> Details
Charles Pandian
Keymaster

I think that you are handling the list in a correct way.

The reason for the problem may be as follows :

If you are creating a list and adding members from one function (or scope) and referring it in another function(or scope), then you may use it as a global variable or
instant variable as follows :



# if you are creating the list here and adding members here

# or any other procedure binded with an agent





Agent/myagent instproc handlelilst {

$self instvar target_list



foreach link $target_list {

puts " link $link is in array"

}



}





proc handlelilst {

global  target_list

foreach link $target_list {

puts " link $link is in array"

}

}

Probably this info will help you to solve the problem.

If you are still having the same problem, then post some meaningful portion of the code or the whole script then only one can identify the error.

#857 April 2, 2010 at 2:08 pm in reply to: list of tcl object
"> Details
Charles Pandian
Keymaster

First install ns2.34 successfully using the installation script as follows :

Change to allinone-2.34/ folder

# ./install

It will install tclcl.

After successful installation, edit the .profile or .bash_profile with settings.
(for example, the following link will provide the profile settings for ns2.28
http://www.projectguideline.com/modules.php?name=News&file=article&sid=9)

After that only tclcl will be in path (the linux shell can recognize it) and you can separately compile ns-2.34 again if necessary as follows :.

allinone-2.34/ns-2.34$ ./configure

#854 March 29, 2010 at 5:35 pm in reply to: Error while using ./configure command while recompiling ns2.
"> Details
Charles Pandian
Keymaster

Yes.
Try using oldest version that can be installed in your machine.

#837 March 29, 2010 at 1:04 pm in reply to: ns2 installation
"> Details
Charles Pandian
Keymaster

Just to make a node to drop a packet, no need to change any routing protocol. You may error model of ns2 to make nodes to randomly drop packets.

#853 March 29, 2010 at 9:16 am in reply to: help me for change code in AODV
"> Details
Charles Pandian
Keymaster

I hope that you can not install ns2.1b9a on Fedora 8 because it will need some lower versions of gcc (and some old version of libraries/header files).

If you try ns2.1b9a on any higher versions of linux, then probably you will see lot of “type errors”

And probably you can con install old versions of linux on new machines (or laptops)

It will need lot of experience for installing a old version of ns2 on high version of linux.
or
It will need lot of experience for installing a old version of linux on a latest computer (for the purpose of installing a old version on ns2).

#835 March 29, 2010 at 8:06 am in reply to: ns2 installation
"> Details
Charles Pandian
Keymaster

try using
Node set multiPath_ 1

instead of
node_ set multiPath_ 1

Use that line just after creating the simulator instance and just before creating any node.
Probably it will work as you expected.

I used in one of my previous work, and it worked correctly.

a similar discussion con be found in the following link.

http://mailman.isi.edu/pipermail/ns-users/1999-November/006175.html

#852 March 28, 2010 at 8:00 pm in reply to: Multipath routing problem
"> Details
Charles Pandian
Keymaster

nrlolsrd seems to be the Naval Research Laboratory’s separate implementation of olsr protocol (not a patch for existing version of olsr of ns2).
protolib seems to be a generic framework for easy development of protocols for ns2 and other platforms.

so I belive that, technicaly both of them can co exist in one instalation of ns2.

#851 March 27, 2010 at 6:53 pm in reply to: regarding installation of nrlolsrd and protolib
"> Details
Charles Pandian
Keymaster

The scripts available from Internet (from previous projects) are used for doing trace analysis in a very generic manner for some specific/ideal scenario/case. If there will be an unusual event happening in the scenario then it will be reflected in the trace file. So the trace analysis script may fail with segmentation fault or divide by zero error or type error.
So just try to change the awk script little bit to handle such unusual situation.

The following link is discussing about another possibility.
http://mailman.isi.edu/pipermail/ns-users/2009-April/065265.html

#849 March 27, 2010 at 6:40 pm in reply to: Problem with simulation of AODV and DSR
"> Details
Charles Pandian
Keymaster

I just checked the tcl example sctipt given in the site
http://www.netmode.ntua.gr/~dbelen/SaTPEP/
(inside the file http://www.netmode.ntua.gr/~dbelen/SaTPEP/satpep-ns-2.27.tar)

The script is getting the loss as a parameter but it is using the parameter to create a folder with that name for redirecting outputs.

The references to that parameter were commented out and the script is using some other loss parameters irrespective of the given input.



###############################################################################

# Loss configuration



set loss [new ErrorModel/List/Seqno]

$loss droplist 150 155 180

$loss flowid 0

$ln_for insert-linkloss $loss



#set loss [new ErrorModel]

#$loss set rate_ $pk_loss_rate

#$loss unit pkt

#$loss ranvar [new RandomVariable/Uniform]

#$loss drop-target [new Agent/Null]

#

#$ln_for insert-linkloss $loss



###############################################################################
#847 March 26, 2010 at 9:15 am in reply to: SatPEP simulation problem
"> Details
Charles Pandian
Keymaster

What kind of filter (with a script like awk /perl) you are using to trace analysis on the generated trace files?

Post the relevant code/script for evaluation.

#846 March 26, 2010 at 7:01 am in reply to: evalvid wireless problem
"> Details
Charles Pandian
Keymaster

The the code for the protocol DV is readily available under ns2 directory. What do you mean by this question?

#818 March 26, 2010 at 6:04 am in reply to: code of distance vector routing alogorithm in ns2
"> Details
Charles Pandian
Keymaster

It is not that much simple.

I think that, you have to write an entire location aware routing protocol to get the direction of the source node which initiated the route discovery. or at least make the source node to send its location inside the packt (the source node should use some (simulated) gps kind of function to determine its location)

#842 March 26, 2010 at 6:02 am in reply to: How to find the direction of packet received or source node?
"> Details
Charles Pandian
Keymaster

It will depend upon the protocol you are using/writing.

For example, if you are going to improve existing version of AODV and make it to behave as Power Aware AODV then you have to do those modifications mainly in aodv.cc and aodv.h and other files in aodv folder of ns2 directory.

If you are writing your own protocol, then you have to write everything in your code and just need to do little modifications on some ns2 files (mostly, packt.h packet.cc, ns-lib.tcl, ns-default.tcl, Makefile etc.,) to integrate your protocol with ns2.

If you just want to change the default values of parameters such as tx power or rx power then you can do it in your tcl simulation code. If you want to change those values permanently, then you may set it in in
ns-allinone-2.xx/ns-2.xx/tcl/lib/ns-default.tcl or the files related with energy model of ns2.

#832 March 26, 2010 at 5:57 am in reply to: can u please help me to set the constraints for energy &
"> Details
Charles Pandian
Keymaster

which version of patch you are using?
which version of ns2 you are using?
what kind of errors you are facing?
Provide a link to get that patch then only others can try it.

#833 March 26, 2010 at 5:43 am in reply to: ns2 installation
"> Details
Charles Pandian
Keymaster

First learn about the ns2’s old and new trace formats.
And then search for suitable strings in the corresponding .cc files.
(carefully examine all the existing ‘fprintf’ commands in those files)

Hopefully, you will get your location to add your own trace events.

#844 March 26, 2010 at 5:40 am in reply to: How are mac802.11 packets traced in cmu-trace.cc?
"> Details
Charles Pandian
Keymaster

Which version of patch you are using?

What kind of experiment you are doing (wired or wireless)?

#845 March 26, 2010 at 5:34 am in reply to: Antnet Protocol for Ad Hoc Network
"> Details
Charles Pandian
Keymaster

In the ns2 world, you can not get anything ready-made most of the times.
All the freely available scrips can be used if you are understanding them.
They will need little modifications to work with your trace files.

As far as a ns2 project is concerned, the trace analysis is the most important part of the project. In other words, it will be the “project” if it is an evaluation type of project.

so doing the project is nothing but learning the trace formats and writing customized scripts for trace analysis.

The following are the some general guidelines of viewing a script which is used for trace analysis :

Are we doing a trace analysis on wired trace or wireless trace format (or any other special non standard format)?

What kind of application event we are interested in and measuring (application layer, transport layer, network layer, mac layer etc.,) ?

What kind of graph we are going to prepare and how (like line chart, bar chart etc.,)?

#843 March 26, 2010 at 5:32 am in reply to: Can anyone help me get awk or, perl parser for simulations?
"> Details
Charles Pandian
Keymaster

After successful installation, edit the .bash_profile with the following values – then only the commands ns nam and xgraph will work) [you can find the file under /home/username/ directory]
Here I am giving the profile setting for ns2.28.
You have to change the version number and folder name according to your installation.





LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/ns-allinone-2.28/otcl-1.9:/ns-allinone-2.28/lib:/root/bin/glnx86/



TCL_LIBRARY=/ns-allinone-2.28/tcl8.4.5/library



PATH=$PATH:$HOME/bin:/ns-allinone-2.28/bin:/ns-allinone-2.28/tcl8.4.5/unix:/ns-allinone-2.28/tk8.4.5/unix



export USERNAME BASH_ENV PATH LD_LIBRARY_PATH TCL_LIBRARY
#840 March 8, 2010 at 12:08 pm in reply to: Executing .tcl
"> Details
Charles Pandian
Keymaster

While the customers trying to transfer the existing fund to their bank accounts, the transfer failed and PayPal charged heavily for that failed transfer.

Few days after the failure of the Paypal to bank transfer, Paypal returned the charges that was originally made during the failed transfer. But still Paypal is not confirming the possibility of Paypal to Indian bank transfer of funds.

#519 February 21, 2010 at 6:05 pm in reply to: Is PayPal Legal in India?
"> Details
Charles Pandian
Keymaster

What do you want to implement and in what protocol?

#830 February 21, 2010 at 6:01 pm in reply to: can i use classifier function in wireless to add my functi
"> Details
Charles Pandian
Keymaster

Of course, the speed of the machine/processor, size of RAM, version of OS – all of them will certainly have an impact on ns2 simulation irrespective of the platform (linux+ns2 / windows+cygwin+ns2). Several years back, while starting my experiments with ns2, I used a Pentium II 233 M.Hz computer with 64 Mb Ram and 2GBHDD(RHL linux 6/7) and was able to run lot of wired and wireless simulation with lot of nodes.

So the speed of a computer will not be a problem most of the times.

For simulating one second of a network scenario, it may take few minuets or few hours for the simulator if the scenario is huge and complex –even with fastest computers. Or it may take less than a second for the simulator, if the scenario is small or we are using a fast computer for the simulation. http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

So, according to my belief and previous experience, most of the times, I realized that the simulation time has been mainly depend upon the complexity of the simulation not the computer.

But, a fast computer will always save lot of time. If you are really getting irritated by the speed of your present computer, then do not go for upgrading your processor or increasing the RAM -Just try for getting a latest computer with multi core processor and highest possible RAM – if you can not spend that much, then just learn to optimize your existing computer for better performance.

Most of the times, the speed of a computer will depend up on the knowledge of person who is using it. so you learn to make it work faster for you.

#826 February 21, 2010 at 5:48 pm in reply to: to install ns2 2.34
"> Details
Charles Pandian
Keymaster

First lean to compile ns2 (I mean – after installing nsallinone package , if everything work good, you can seperately compile ns2 from
./ns-allinone-2.xx/ns-2.xx folder)

do the following

if you want to include some function in aodv, then add them inside aodv.cc and aodv.h according to the syntax (normal c++ way).

#cd ns-allinone-2.xx/ns-2.xx
#make clean (sometimes, it is not necessary)
#make
#make install (sometimes, it is not necessary)

If yo see any compile time error, correct them.

and do the make again.

Thats all.
If your function is logically correct, then it will work.

#829 February 13, 2010 at 12:48 pm in reply to: how include the c code in ns2 coding
"> Details
Charles Pandian
Keymaster

I forgot to tell one thing – The old version of cygwin can not be installed on 👿 Vista/Windows7. The very latest versions of cygwin can only (said to) be installed on Vista/Windows7 😈 (with minimum problems).

Old version of cygwin can be installed on Windows XP without any problems. and even, almost all the versions of ns2 can be made to work on it 😀 .

#824 February 6, 2010 at 3:30 pm in reply to: to install ns2 2.34
"> Details
Charles Pandian
Keymaster

Generally a patch written for ns2 will work on any installation 💡 (both ns2 on linux or ns2 on cygwin). But some times you may face some type errors or deprecated function errors during compiling ns2 after patching it. Most of the times, these kind of errors can be rectifies (using type casting or some simple modification in code) – The most of these errors will mainly depend up on the version of gcc compiler in that platform. It means if you are facing such error while patching ns2 on a particular version of linux then probably you will not have the same error on some higher of lower version of the same linux varient 🙄 (or different version of cygwin – on windows). [Different version of linux/cygwin will packaged with different version of gcc ]

– I think I have confused you too much 😀 – But in fact this is the main problem people are facing while struggling with ns2. Most of the times they fail 😥 because they never understand the thing which I have explained in the above paragraph.

#823 February 6, 2010 at 3:23 pm in reply to: to install ns2 2.34
"> Details
Charles Pandian
Keymaster

Generally, the installation of higher versions of ns2 ( >= 2.28 ) on cygwin will be very much easier than installing it on linux.

#821 February 6, 2010 at 5:21 am in reply to: to install ns2 2.34
"> Details
Charles Pandian
Keymaster

There is a multicast extension of AODV called MAODV available for ns2. You can get the patch and experiment with wireless multicast routing.

#820 February 4, 2010 at 5:13 am in reply to: wireless multicasting
"> Details
Charles Pandian
Keymaster

Since the versoin of MAODV you are using is meant for ns 2.26, running it on ns2.26 only be the easy choice. Further, ns 2.26 can be installed on old Red Hat linux versions without any problem. – Now you are spending your efforts in correct direction. -It is the best way of doing this.

You can possibly compile ns2.26 successfully on resent versions of Fedora or Ubuntu just by using old version of gcc and correcting some errors. – But I will not prefer it.

You may use a resent version of ns2 on Fedora or Ubuntu – but in that case, the MAODV patch will fail. You may compile ns2 with MAODV by applying the MAODV patch manually (if possible) and correct some errors. – But certainly it will require lot of effort.

#816 January 25, 2010 at 9:51 am in reply to: MAODV on newer versions of NS and Linux
"> Details
Charles Pandian
Keymaster

As far as I tested, RHL and RHL variants like Fedora will not have much problems during installing ns2. But, the successful installation of ns2 on a particular version of linux will depend on the version of several components like gcc, X lib and other libraries.

#806 January 14, 2010 at 4:21 pm in reply to: Help regarding LEACH clustering Protocol
"> Details
Charles Pandian
Keymaster

First read the following post
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

and then read the following document
http://masimum.inf.um.es/nsrt-howto/pdf/nsrt-howto.pdf

Yes. it may take years

#812 January 13, 2010 at 6:28 am in reply to: guide me for implementing below algorithm in NS2
"> Details
Charles Pandian
Keymaster

First read the following post
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&t=13

and then read the following document
http://masimum.inf.um.es/nsrt-howto/pdf/nsrt-howto.pdf

#811 January 13, 2010 at 6:28 am in reply to: guide me for implementing below algorithm in NS2
"> Details
Charles Pandian
Keymaster

Most probably, most of the existing wifi extensions of ns2 will not work with ns2.34. So decide which version of wifi extensions to use and then decide the suitable lower version of ns2 based on it.

#809 January 13, 2010 at 6:21 am in reply to: ns2 project
"> Details
Charles Pandian
Keymaster

NSG is just a JAVA based software. So only you need to worry about the working environment java – not the operating system.

You can get it from here (I think you already know this)
http://sites.google.com/site/pengjungwu/nsg

#807 January 8, 2010 at 5:08 am in reply to: Network Scenario Generator
"> Details
Charles Pandian
Keymaster

0.05 means 5% .
That is, only 5% of the node will behave as cluster head. It is a belief that the 5% nodes will have distributed locations to cover overall network.
If you use 0.25, then 25% of the nodes will behave as cluster head.

#804 January 8, 2010 at 5:00 am in reply to: Help regarding LEACH clustering Protocol
"> Details
Charles Pandian
Keymaster

This problem may be due to the gcc version of Fedora 9. Try installing lower version of gcc.

OR

Try installing higher versions of ns2 on Fedora 9

OR

Try ns2 on lower version of Fedora.

ns2 can be installed on RedHat Linux (I tried only on old versions of RHL) without any gcc related issues.

#803 January 1, 2010 at 7:27 pm in reply to: NS2 Installation Problem
"> Details
Charles Pandian
Keymaster

Old versions of network simulators will require patching before the installation.

So before run the ./install, you have to apply suitable patch for that version (ns2.1b8a).

Try to find a patch file called ns2.1cygwin.patch.
you may make ns2.1b8a to work by applying this patch and with some little further modifications.

Once I made it (ns-allinone-2.1b9a) work on cygwin.

#802 December 31, 2009 at 8:38 am in reply to: please help
"> Details
Charles Pandian
Keymaster

Probably it means that you have not successfully patched it.
Do a clean installation of a suitable version of ns2 and watch the error while patching and try to correct it manually.

#800 December 24, 2009 at 3:51 pm in reply to: help
"> Details
Charles Pandian
Keymaster

Post the script which you are using to calculate the throughput and portion of of the trace file. without this two no one can find your mistake.

#799 December 21, 2009 at 1:43 pm in reply to: zrp in ns 2.30
"> Details
Charles Pandian
Keymaster

@zabinskia wrote:

How can I change the PayPal account linked to my ebay sellers account? My buyer has submitted payment, but it went to an e-mail address that is not linked to any paypal account. I do have a paypal account, but don’t know how to verify my Ebay sellers account is linked to it.

First, send an apology mail to your buyer(if you know his mail id).

I think you (preferably your buyer) have to make a complaint to paypal regarding this wrong payment and claim that money back to your buyer’s account.

You may refer the following links for more information.

http://pages.ebay.com/help/pay/accepted-payment-methods.html#paypal

https://www.paypal.com/us/cgi-bin/webscr?cmd=xpt/Marketing/securitycenter/buy/Protection-outside

Before starting real business transactions, you have to experiment them with your own accounts. Paypal is providing some virtual environment to test these kind of transactions with some dummy accounts. Just explore papal’s site you will find it.

#798 December 14, 2009 at 8:02 am in reply to: PayPal account linked to my ebay
"> Details
Charles Pandian
Keymaster

😀 As far as the HOTness of ns2 is concerned, Red Hat is the better Hat than Fedora (Fedora- A hat made of felt with a creased crown)

#562 December 9, 2009 at 11:20 am in reply to: Get tcl error during installing ns2.29 on Fedora and ubuntu
"> Details
Charles Pandian
Keymaster

You may get the idea by exploring OLSR code.
Refer this link for installing OLSR in ns2.
http://www.projectguideline.com/modules.php?name=News&file=article&sid=15

#796 November 30, 2009 at 4:04 am in reply to: Link state with Hello packets
"> Details
Charles Pandian
Keymaster

The following document explains a procedure
http://www.cs.binghamton.edu/~nael/cs528/lqsr.pdf
You may already saw this one. But it will be the only available starting point for your implementation.

#794 November 26, 2009 at 7:29 am in reply to: LQSR in ns2
"> Details
Charles Pandian
Keymaster

yes mr charles,, thank you for the advice,, but i need the right scenario simulation to see the behavior of scheduling algorithm role’s in wimax,, do u had any suggestion? what is the meaning of worst case scenario simulation ?

I think, NIST WiMax extension model will be the most suitable one for your experiments since there seems a real implementation of a scheduler (BE -based) in it.

I think that a “worst case scenario” simulation is one in which some of the input parameters of the simulations are not taken in to account and hence that may lead inaccurate or approximate results.- If I am wrong some one may define it in a better way.

#770 November 26, 2009 at 6:59 am in reply to: WiMAX scheduling scenario ns2
"> Details
Charles Pandian
Keymaster

Very long time ago, I made the ns2.26 version of GnutellaSim to work on cygwin and tested with some preliminary scenario. But its implementation seems to be too complicated.

#765 November 24, 2009 at 8:44 pm in reply to: Gnutellasim
"> Details
Charles Pandian
Keymaster

Since the theoretical maximum raw data rate of 802.11b seems to be 11 Mbit/s, the default parameters possibly try to set that value. But, the parameters you mentioned here is not setting it.

For example in 802.11, the following line will set its default data rate.
Mac/802_11 set PLCPDataRate_ 1.0e6 ;# 1Mbps

(IEEE 802.11 – The WLAN standard was original 1 Mbit/s and 2 Mbit/s)

#792 November 24, 2009 at 8:29 pm in reply to: setting data rate for ieee802.11p
"> Details
Charles Pandian
Keymaster

Upto now, none of the existing WiMax extensions having any complete scheduler implementation. I hope that, Best Effort based scheduler is available in one of the available extension. But it is also in very preliminary stage. So you have to wait for some time (for some expert to do it) or you have to ‘bell the cat’ on your own.

#768 November 24, 2009 at 8:00 pm in reply to: WiMAX scheduling scenario ns2
"> Details
Charles Pandian
Keymaster

If you are facing problems related with installation or facing errors during running the simulation, then it may be due to the ns2 version you are using. That extension can be used only on a particular old version of ns2.

#784 November 24, 2009 at 7:53 pm in reply to: MIPv6 simulation (Using Mobiwan Tool)
"> Details
Charles Pandian
Keymaster

Give little more explanation about your requirement/scenario

#787 November 24, 2009 at 7:49 pm in reply to: how to pause ns2 scheduler from c++ code?
"> Details
Charles Pandian
Keymaster

NS by Example by Jae Chung and Mark Claypool http://nile.wpi.edu/NS/
explains RED Queue Monitoring through an Example.

#764 November 24, 2009 at 7:45 pm in reply to: IFQ size in packets
"> Details
Charles Pandian
Keymaster

After doing necessary modifications, you have to recompile ns2 using makefile as follows :

# cd ns-allinone-2.28/ns-2.28
# make
# make install
(the last step may be optional)

This will compile the code and generate new ns binary with your proposed modification.

Still ff you are not seeing any difference, then it may due to your modification which may probably has no impact on its performance.

#790 November 24, 2009 at 7:39 pm in reply to: TDMA functioning for Ad hoc Networks
"> Details
Charles Pandian
Keymaster

The section 4.1 of original technical document “rsvpns.pdf” by Marc Greis explains the procedure. It says, that the README.RSVP document which comes along with the patch has the detailed patching procedure.

Refer the following document.

http://www.redes.unb.br/material/Topicos_Redes_Comunicacao/rsvpns.pdf

#788 November 23, 2009 at 10:44 am in reply to: RSVP patch.
"> Details
Charles Pandian
Keymaster

Some very old versions of ns2 (like ns2.1xx) was available for windows compilation (using Microsoft Visual C++).

But any version of ns2 can be compiled under Cygwin Unix/Linux Emulator for windows.

But the final binaries (like ns.exe or nam.exe or xgraph.exe) can be used under cygwin environment only.

Even the compiled binaries can be made to run in a stand alone fashion on another windows systems by using suitable dependencies of cygwin binaries (.dll files).

#789 November 19, 2009 at 7:00 am in reply to: converting .out in Fedora Linux to .exe in windows
"> Details
Charles Pandian
Keymaster

Read Chapter “VII. A new protocol for ns” of Marc Greis’ Tutorial.
http://www.isi.edu/nsnam/ns/tutorial/
This is explaining a simple case.

Read the following.
http://masimum.inf.um.es/nsrt-howto/pdf/nsrt-howto.pdf
It is explaining the creation of a protocol in detail.

#785 November 18, 2009 at 5:56 pm in reply to: how to create my own packet in ns2?
"> Details
Charles Pandian
Keymaster

Just read the patching procedure explained in the above mentioned link
(see my previous post).
It will solve the problem.

As far as I know, the TORA implementation in almost all the higher versions of ns2 (possibly >ns2.26) will not work by default. You can make it work after the above mentioned patch operation.

But, according to my previous experiments with almost all the versions of ns2, I manage to make it worked but TORA never gave any better results as explained in previous comparisons of research papers and theories.

#759 November 18, 2009 at 7:52 am in reply to: Tora Protocol is not working under 2.34 version
"> Details
Charles Pandian
Keymaster

It seems that, up to now, no one has provided HMRSVP patch for NS2 as a free download.

So you have to write HMRSVP for ns2 on your own.

If anyone finds a similar work for ns2, then kindly post it in this thread.

#782 November 12, 2009 at 5:25 am in reply to: HMRSVP patch.
"> Details
Charles Pandian
Keymaster

It seems that all the original links for downloading windows version of it has been removed for some unknown reason (The servers maybe temporarily down 😀 or it may will not be free any further 😥 – or it may be removed due to any copyright issues 👿 – If someone is aware of its status, then kindly post it in this thread)

But it seems that, a linux version of it is still available at the following link 💡 .
http://www.angelfire.com/al4/esorkor/trgraph155clinux.tar.gz
I found it in the following website
http://www.angelfire.com/al4/esorkor/

#780 November 12, 2009 at 5:14 am in reply to: Trace Graph.
"> Details
Charles Pandian
Keymaster

I hope that, you should look in to the code of Propagation Models of ns2

#777 November 4, 2009 at 5:27 pm in reply to: hi …
"> Details
Charles Pandian
Keymaster

Sorry for not answering your question.

Why are you needing such a complex and huge IDE such as Eclipse to compile ns2 ?
./install in a terminal window will do the same thing in more efficient and obvious way.

If you want syntax highlight you may use a simple and good text editor like SciTE to edit the .cc and .h files.
If you want to debug your code with an IDE, then in that case also, we can trace the error just by using some “printf” in appropriate places. (this is the simple way which I am using for debugging 🙂 )

As far as I know, while developing a new protocol the roll of an IDE is very very insignificant.

Once I used a light weight IDE called “Geany IDE” to compile ns2. It was good and really fast.

Kindly explain the need for using Eclipse to compile ns2.

Any way I hope that you are following a procedure like the following one
http://blog.xizi.com/?uid-15519-action-viewspace-itemid-44474

#778 November 4, 2009 at 5:11 pm in reply to: Problem compiling ns-2.34 with eclipse Galileo under ubuntu
"> Details
Charles Pandian
Keymaster

@Sherry wrote:

Hi all
>
> I need help in running a tcl scripts and testing the performance of the
> MAODV protocol in ns2 version 2.33.
> I know that for most ns-users this is pretty prementive, but for me as I’m
> new to ns2 is a hard task. I followed the thesis ” MAODV Implementation for
> NS-2.26″ by Yufang Zhu and Thomas Kunz. step bu step and finally got this
> error message when I run ns ns.tcl in the X window:
> *****************************
> num_nodes is set 50
> INITIALIZE THE LIST xlisthead
>
> (-o17 cmd line 1)
> invoked from within
> “-o17 and port-dmux-o26”
> invoked from within
> “catch “$self cmd $args”ret”
> invoked from within
> “if [catch “$self cmd $arg”ret] {
> set cls [$self info class]
> global errorInfo
> set savedInfo $errorInfo
> error”error when calling class $cls:$args” $…”
> (procedure “-o17” line2)
> (splitObject unknown line2
> **************************
> Please help me in trying to run the script
> Thank u in advance

It simply means that you not yet successfully did the MAODV patch.

#661 November 2, 2009 at 4:22 pm in reply to: MAODV implementation Problem on ns2.33
"> Details
Charles Pandian
Keymaster

@Sherry wrote:

Hi all
>
> I need help in running a tcl scripts and testing the performance of the
> MAODV protocol in ns2 version 2.33.
> I know that for most ns-users this is pretty prementive, but for me as I’m
> new to ns2 is a hard task. I followed the thesis ” MAODV Implementation for
> NS-2.26″ by Yufang Zhu and Thomas Kunz. step bu step and finally got this
> error message when I run ns ns.tcl in the X window:
> *****************************
> num_nodes is set 50
> INITIALIZE THE LIST xlisthead
>
> (-o17 cmd line 1)
> invoked from within
> “-o17 and port-dmux-o26”
> invoked from within
> “catch “$self cmd $args”ret”
> invoked from within
> “if [catch “$self cmd $arg”ret] {
> set cls [$self info class]
> global errorInfo
> set savedInfo $errorInfo
> error”error when calling class $cls:$args” $…”
> (procedure “-o17” line2)
> (splitObject unknown line2
> **************************
> Please help me in trying to run the script
> Thank u in advance

This error means that you not yet successfully did the MAODV patch.

#660 November 2, 2009 at 4:22 pm in reply to: MAODV implementation Problem on ns2.33
"> Details
Charles Pandian
Keymaster

Almost all the protocol elements of ns2 (or agents) will contain two important functions. One is “recv” and the another is “send”. So, in your protocol agent’s code, you have to find a appropriate location (probably, in recv function) to implement the required functionality.

#767 November 2, 2009 at 4:19 pm in reply to: FullTcp upcall
"> Details
Charles Pandian
Keymaster

First you may experiment with tcl level parameters of the protocol (in side your simulation script) and try to create awk scripts for trace analysis (for delay ,throughput ,jitter). After successful results or creation of some graphs, then you may try to alter/improve the c+ level ns2 code.

#775 November 2, 2009 at 4:10 pm in reply to: hi ns2 problem..enviromental conditions.
"> Details
Charles Pandian
Keymaster

I hope that, an old verions of RSVP patch is available at
http://www.netlab.tkk.fi/tutkimus/ironet/ns2/rsvp.zip
You may start your work based on this.

#773 November 2, 2009 at 4:05 pm in reply to: MRSVP & M-YESSIR
"> Details
Charles Pandian
Keymaster

I found a solution in a previous post at
http://mailman.isi.edu/pipermail/ns-developers/2007-March.txt



/usr/bin/ld: Undefined symbols:

_TkCreateXEventSource

collect2: ld returned 1 exit status

make: *** [nam] Error 1



The issue is still in tcl/tk shipped by Apple: nam Makefile tries to

link with tcl/tk in /usr/lib. I sorted out this way: first, I created

a link to the core osx library inside nam directory:



$: cd nam-1.13

$: ln -s /System/Library/Frameworks/CoreFoundation.framework/Versions/

A/CoreFoundation libhack.dylib



Then I linked nam with static tcl/tk in ns-allinone:



$: g++ -o nam 

tkcompat.o 

tkUnixInit.o 

xwd.o 

netview.o 

netmodel.o 

edge.o 

packet.o 

node.o 

main.o 

trace.o 

queue.o 

drop.o 

animation.o 

agent.o 

feature.o 

route.o 

transform.o 

paint.o 

state.o 

monitor.o 

anetmodel.o 

random.o 

rng.o 

view.o 

graphview.o 

netgraph.o 

tracehook.o 

lan.o 

psview.o 

group.o 

editview.o 

tag.o 

address.o 

animator.o 

wnetmodel.o 

nam_stream.o 

enetmodel.o 

testview.o 

parser.o 

trafficsource.o 

lossmodel.o 

queuehandle.o 

gen/version.o 

gen/nam_tcl.o  

../tcl8.4.14/unix/libtcl8.4.a 

../tk8.4.14/unix/libtk8.4.a 

../tcl8.4.14/unix/libtclstub8.4.a 

-L/Users/gringoli/Work/ns-tutto/ns-allinone-2.31/tclcl-1.19 -ltclcl 

-L/Users/gringoli/Work/ns-tutto/ns-allinone-2.31/otcl-1.13 -lotcl 

-L/Users/gringoli/Work/ns-tutto/ns-allinone-2.31/lib  

-L/Users/gringoli/Work/ns-tutto/ns-allinone-2.31/lib 

-L/usr/lib -lz -L/usr/X11R6/lib -lXext -lX11 -lm -L. -lhack



Regards,

FG

It seems to be a solution for the problem with tcl/tk shipped by Apple.
After experimenting with it, kindly post the result here.

#748 August 17, 2009 at 9:25 pm in reply to: ns-2 installation on Mac OS X (Leopard)
"> Details
Charles Pandian
Keymaster

There is no simple or readymade solution to your problem.
You have to understand the AODV code which is available in the directory ns-allinone-2.xxns-2.xxaodv

After little bit of effort, you may add the required functionality in the existing AODV code.

First study the following howto
http://masimum.inf.um.es/nsrt-howto/pdf/nsrt-howto.pdf

#751 August 14, 2009 at 6:22 pm in reply to: hashtable implement in aodv
"> Details
Charles Pandian
Keymaster

If possible, upload your tcl script. Then only some one in this forum may provide you a solution.

#752 August 14, 2009 at 6:17 pm in reply to: 2 mobile node can’t send the TCP packet at same time
"> Details
Charles Pandian
Keymaster

The default version of TORA in earlier version of ns2 requires little patching for working.

The following links provide that information.

http://www.gidforums.com/t-19724.html
http://mailman.isi.edu/pipermail/ns-developers/2006-March/002250.html

I followed the procedure and made TORA work on ns2.26.

This patch may solve your problem

#756 August 14, 2009 at 6:14 pm in reply to: Tora Protocol is not working under 2.34 version
"> Details
Charles Pandian
Keymaster

The following link provides a instructions for patching ns2 for IEEE 802.16d

http://cng1.iet.unipi.it/wiki/index.php/Ns2mesh80216

Hopefully the procedure will work on ns-allinone-2.33

#760 August 14, 2009 at 6:09 pm in reply to: Patch ns2mesh80216-2.33-081113
"> Details
Charles Pandian
Keymaster

The following link provides a proceedure for installing ns2.31 on OS X 10.5.5

http://sites.google.com/site/pengjungwu/install-ns2-31-on-osx

Hopefully the same procedure will work for you.

#746 August 14, 2009 at 6:06 pm in reply to: ns-2 installation on Mac OS X (Leopard)
"> Details
Charles Pandian
Keymaster

These kind of installation problems are mainly due the version of gcc which is installed on the linux os.

First apply suitable gcc patch for your version of ns2, for your version of gcc, before starting the installation.

If you are getting same kind of errors, then try to replace the current version of gcc with version lower than gcc v3.3

Hopefully it will solve the problem.

#741 August 5, 2009 at 8:16 am in reply to: ns-2 installation
"> Details
Charles Pandian
Keymaster

These kind of installation problems are mainly due the version of gcc which is installed on the linux os.

First apply suitable gcc patch for your version of ns2, for your version of gcc, before starting the installation.

If you are getting same kind of errors, then try to replace the current version of gcc with version lower than gcc v3.3

Hopefully it will solve the problem.

#560 August 5, 2009 at 8:14 am in reply to: Get tcl error during installing ns2.29 on Fedora and ubuntu
"> Details
Charles Pandian
Keymaster

sorry your question is not clear.

Are you talking about some kind of table management for handling some global variables in a c++ level agent?

#740 July 5, 2009 at 4:54 am in reply to: how to implement Lookup Table in NS-2
"> Details
Charles Pandian
Keymaster

refer the following answer provided in the following link.
http://www.projectguideline.com/modules.php?name=Forums&file=viewtopic&p=452#452

@CharlesPandian wrote:

Try to eanable the new trace format of ns2 before simulation.

While doing trace analysis, filter sent or recieved packet from suitable layer (MAC / Routing / AGT). This will avoid false counts.

Still, there are chances for getting high received count due to duplicates. We can isolate duplicates them using the unique packet ID.

#728 May 27, 2009 at 2:12 pm in reply to: number of packets received is more than the packets sent
"> Details
Charles Pandian
Keymaster

It is a very common error any one will face during developing a new agent or protocol for ns2.
It means you somewhere did a mistake in while patching ns2 files.

First, Understand the basic tutorials of ns2 for creating a new agent for ns2.

Start with “chapter VII -A new protocol for ns” of Marc Greis’ Tutorial

http://www.isi.edu/nsnam/ns/tutorial/

#733 May 27, 2009 at 2:00 pm in reply to: Error with new agent
"> Details
Charles Pandian
Keymaster

There are lot of such rate based congestion control protocols available on internet. Just use google to find a suitable one with respect to your requirement.

Start searching from this two links
http://netlab.caltech.edu/projects/ns2tcplinux/ns2linux-2.29-linux-2.6.16/
http://nsnam.isi.edu/nsnam/index.php/Contributed_Code

#736 May 27, 2009 at 1:50 pm in reply to: WCCP or RBCC source code
"> Details
Charles Pandian
Keymaster

Try to eanable the new trace format of ns2 before simulation.

While doing trace analysis, filter sent or recieved packet from suitable layer (MAC / Routing / AGT). This will avoid false counts.

#737 May 27, 2009 at 1:45 pm in reply to: Bad DSR performance due to packet drop at IFQ
"> Details
Charles Pandian
Keymaster

Learn about Energy Model from ns2 manual. While using Energy Model in a ns2 simulation, the following line can be used to get Energy of the node

set RemainingEnergy [$node(1) energy]

The variable RemainingEnergy will contain the value.

#726 April 18, 2009 at 5:07 pm in reply to: measuring battery power of nodes in ns2
"> Details
Charles Pandian
Keymaster

You have to set tcl path. If you are installing ns2-allinone by running the “install script” then it may refer the path of newly compilied tcl version correctly, and pass it to the other installation steps. But if you are directly compiling the ns-2 only part, then you need to set tcl path correctly before running configure and make.

#723 March 21, 2009 at 7:36 pm in reply to: problem in installing ns-2.30 in fedora
"> Details
Charles Pandian
Keymaster

As far as ns2 is concerned, generally, the tcl file will contain the main simulation code. Sometimes, additional tcl files also can be used from the main simulation script.

The main ns2 simulator code itself was written in .cc, .h and tcl files. It means an ns2 object can be a c++ level implementation or a tcl level implementation.

Generally while doing a project on ns2, some of the packet level functionalities will be implemented in c++ code and those lower level function or classes will called/used from a tcl script.

Using an output file with extension .tr is a convention in ns2 – generally, those files are used to log some  event traces for further trace analysis.

#715 February 26, 2009 at 12:03 pm in reply to: difference between .tcl,.cc,.h,.tr
"> Details
Charles Pandian
Keymaster

AODV code is already available in standard ns2 distribution.(under ns-allinone-2.33ns-2.33aodv in ns version 2.33 )

#712 February 20, 2009 at 4:18 pm in reply to: help for aodv code on ns 2
"> Details
Charles Pandian
Keymaster

The this question was already answered. Search this forum; you will find the answer.

#698 January 29, 2009 at 5:26 am in reply to: sending data
"> Details
Charles Pandian
Keymaster

Since the second script is including the first one (using source command), you have to run the second script only. But the script is expecting some command line parameters. So give them while running the code like

#ns mycode.tcl parameter1 parameter2 etc.,

#685 January 29, 2009 at 5:22 am in reply to: how to run tow tcl files
"> Details
Charles Pandian
Keymaster

Yes you can. But not with the default TCP/UDP agents available in ns2.

#700 January 29, 2009 at 5:17 am in reply to: data sending in ns2
"> Details
Charles Pandian
Keymaster

Since the patch is for ns-2.30, first try the patch after installing a fresh copy of ns-2.30.

#710 January 29, 2009 at 5:16 am in reply to: umts in ns2
"> Details
Charles Pandian
Keymaster

Technically , there is no difference in running a ns2 code on windows(with cygwin) as well as Linux.

If your code is running some external commands using shell, then it will differ in both environments.

#701 January 29, 2009 at 5:13 am in reply to: transfer commands from ns2 on cygwin to ns2 on fedora
"> Details
Charles Pandian
Keymaster

After validating your paypal account, you can directly transfer the money from your paypal account to some of (all?) the Indian bank accounts. But for validating your paypal account, you should have a valid credit card.

#702 January 29, 2009 at 5:09 am in reply to: Withdrawing money thru foriegn transaction using paypal
"> Details
Charles Pandian
Keymaster

It is already available in ns2. What you have to do is, you should learn to use it.

#703 January 29, 2009 at 5:04 am in reply to: GEAR ns-2 code
"> Details
Charles Pandian
Keymaster

There are some awk scripts freely available for doing trace analysis. Just use google to find suitable scripts for your requirement. (use suitable keywords like trace analysis, ns2 .awk etc)

#705 January 29, 2009 at 5:01 am in reply to: Throughput and collision measurement in 802.11 AP-mode
"> Details
Charles Pandian
Keymaster

You have to connect them separately as usual.
But you can minimize the code by doing this in a for loop.

#708 January 29, 2009 at 4:56 am in reply to: Connecting a number of nodes to a one node
"> Details
Charles Pandian
Keymaster

As far as I know, there is no freely available code extension for doing content based routing in ns2.

But there are lot of papers related with this routing protocol
http://www.google.co.in/search?hl=en&q=content+based+routing+ns2+filetype%3Apdf&btnG=Search&meta=

So if you are doing a research/project related with CBR then you have to develop the routing agent on your own starting from single line.

If you are trying to develop the protocol for wired network, then it will be somewhat easier to implement.

#709 January 29, 2009 at 4:54 am in reply to: tool for content based routing of xml publish /subscribe p2p
"> Details
Charles Pandian
Keymaster

The answer is not simple. Once I did that for a research project (It may be a indirect way and there may be some other simple way of doing the same).

First, you have to learn the basic idea from the above said link.

I can not upload the full code due to some privacy/copyright reasons.

You have to do the necessary coding in a c++ level agent and pass the values through a tcl function. The tcl function (AddNeighborToList) will populate a tcl list type variable.

The main part of C++ level agent code.



void MyAgent::FillNeighborList()

{

Tcl& tcl = Tcl::instance();



int TotalNodes=God::instance()->nodes();



for(int i=0;i
{

if(God::instance()->IsNeighbor(i , here_.addr_))

{

tcl.evalf("%s AddNeighborToList %d ",name(), i);

}

}

}

The tcl code which is actually populating the neighbor list “NeighborList_”



Agent/MyTclAgent instproc AddNeighborToList {nodeid} {

$self instvar NeighborList_

lappend NeighborList_ $nodeid

}
#696 September 21, 2008 at 9:55 pm in reply to: God::instance()->IsNeighbor(i , j)
"> Details
Charles Pandian
Keymaster

I hope that you can not call
God::instance()->IsNeighbor from tcl script since it is a c++ function.

Reply to this post if I am wrong.

#694 September 21, 2008 at 10:09 am in reply to: God::instance()->IsNeighbor(i , j)
"> Details
Charles Pandian
Keymaster

If you are talking about wireless scenario, then you can not detect neighbors of a node like this.

Try the following link.
[url]
http://www.cse.msu.edu/~wangbo1/ns2/nshowto2.html
[/url]

#691 September 19, 2008 at 2:58 pm in reply to: God::instance()->IsNeighbor(i , j)
"> Details
Charles Pandian
Keymaster

Try to run the script as follows :

#ns pbcast_sim.tcl -prob 50

or

ns pbcast_sim.tcl -prob 0.5

#686 September 11, 2008 at 2:38 am in reply to: how to run two tcl files
"> Details
Charles Pandian
Keymaster

The Problem may be related with the following line.

$ns_ connect $proto0 $proto1

#678 August 21, 2008 at 4:03 pm in reply to: Implementing a New Manet Unicast Routing Protocol in NS2
"> Details
Charles Pandian
Keymaster

From which article you took the code?
How you implemented Agent/Protoname?
and what it will do?

#676 August 21, 2008 at 2:36 pm in reply to: Implementing a New Manet Unicast Routing Protocol in NS2
"> Details
Charles Pandian
Keymaster

If it is a wired network, then you may easily resolve neighbor list.

But in MANET, resolving neighbor list is not that much simple.

You can prepare a neighbor list by sending one hop hello broadcast and receiving reply messages of neighbor nodes.

You can directly derive the neighbors from global ns environment using “god”. You may refer the following link for how to use “god” for deriving neighbor list in MANET. But it will also require good knowledge about compiling ns2.

http://www.cse.msu.edu/~wangbo1/ns2/nshowto2.html

#674 August 20, 2008 at 4:51 pm in reply to: how to set the probabilty by useing neighbors variables
"> Details
Charles Pandian
Keymaster

A simple and crude technique 😆 of implementing any kind of new agent in ns2 is using searching for keywords in entire ns2 directory structure.

For example, if you want to implement a protocol similar to AODV, then search for suitable keyword like “aodv” in entire ns2 directory structure 💡 and find the files containing code for “AODV”. After understanding those files, then clone entire file (aodv.cc, aodv.h etc.,) and rename them according to new protocol or add suitable things in some of the files (like commonpacket.h, tracetrace.cc etc)

Once I used the same technique to clone AODV to implement a new improved version of AODV protocol (it almost consumed a week 🙄 to make the new protocol to run correctly.

#673 August 16, 2008 at 3:51 am in reply to: adding new agent
"> Details
Charles Pandian
Keymaster

I can not understand what will be the ‘good’ parameters because, generally we will set parameters to simulate a network most accurately according to it’s application.

#671 August 12, 2008 at 7:12 am in reply to: 802.11 simulation parameters
"> Details
Charles Pandian
Keymaster

The function TrustManager::initNewTrustValues(Path&, int) should have a prototype in the corresponding header file.

Some times, even if you are having prototype defined in header file, you may see this error due to type mismatch in arguments or return type. If it is the cause of the error then use type cast. (Some version of compiler can handle this without type casting. So you will not get such errors in different version of linux/gcc)

#672 August 12, 2008 at 6:58 am in reply to: (.text+0x6b13): undefined reference to function_name
"> Details
Charles Pandian
Keymaster

Ns2 is my choice. Because, it is the most widely used simulator by research community.

#664 August 11, 2008 at 4:45 am in reply to: suggestion on MANET clustering techniques
"> Details
Charles Pandian
Keymaster

In a typical awk script I saw the following line for throughput calculation.

printf(“nAverage Throughput[kbps] = %.2f “,(recvdSize/(stopTime-startTime))*(8/1000))

I could not understand your formula “u*8*500/$2”.

#667 August 8, 2008 at 1:37 pm in reply to: Problem in mobility module ns-2.29 developed by NIST
"> Details
Charles Pandian
Keymaster

There are several reasons to end with Segmentation fault. The most common ones are, due to wrong format strings in fprintf command. (for example, if you try to print a string which is not null terminated, then you may end with such error)

First check the fscanf or fprintf commands (if any)

#670 August 7, 2008 at 7:04 pm in reply to: Segmentation fault
"> Details
Charles Pandian
Keymaster

The error may be in the script you are using to calculate throughput. Explain the network topology and how you are calculating throughput.

#665 August 7, 2008 at 5:33 am in reply to: Problem in mobility module ns-2.29 developed by NIST
"> Details
Charles Pandian
Keymaster

Yes, electing a cluster head based on mobility and remaining battery power will be a good option. You may use voronoi diagram based optimization techniques for electing the cluster head and you may use classical clustering techniques to cluster the nodes.

#662 August 7, 2008 at 4:29 am in reply to: suggestion on MANET clustering techniques
"> Details
Charles Pandian
Keymaster

It will depend up on the version of ns2 as well as the version of linux/unix variant

If you need to use old version of ns2 (like ns2.1b) for a specific kind of project, then I prefer some old Red Hat Linux (7 – 9 )

If you want to try a new version of ns2, then better try it on Fedora.

#591 July 30, 2008 at 4:31 pm in reply to: Fedora Core or Redhat?
"> Details
Charles Pandian
Keymaster

I think that your are running a wrong installation script.

Just explain the procedure (which you are trying) step by step.

#649 July 22, 2008 at 5:45 pm in reply to: installation problem
"> Details
Charles Pandian
Keymaster

First try to reduce the distance between MN and HA and run the simulation.

#646 July 22, 2008 at 10:08 am in reply to: Difficulties in wireless simulation
"> Details
Charles Pandian
Keymaster

I hope that this problem is not related with ns2 installation. Try to install ns2 in some other version of linux or on a clean installation of cygwin.

#648 July 22, 2008 at 9:41 am in reply to: Wu Xiuchaoe error model
"> Details
Charles Pandian
Keymaster

What kind of result you are getting? And what kind of result you are expecting? – Just explain briefly.

What version of ns you are using?

#644 July 22, 2008 at 7:53 am in reply to: Difficulties in wireless simulation
"> Details
Charles Pandian
Keymaster

There is no need to do any special kind of measurement for AP node. There are standard ways to do this kind of measurements through trace analysis.

Learn trace analysis of ns2 trace file.

Read the turorial called “NS by Example” by Jae Chung and Mark Claypool which is available at the following link.

http://nile.wpi.edu/NS/
http://nile.wpi.edu/NS/analysis.html

It will give you brief introduction to what is trace analysis.

You may lean much from similar online tutorials/web pages.

#641 July 22, 2008 at 4:19 am in reply to: Calculate traffic load at Access Point
"> Details
Charles Pandian
Keymaster

There seems to be some implementations in C/C++ (not for ns2).

There is a possibility of adopting one of such implementation and develop the code for ns2.
.
In the following link, there is one such ‘C’ implementation of “Gossip protocol”

http://www.google.co.in/codesearch/url?hl=en&ct=rp&url=http://www.angstrom-distribution.org/unstable/sources/gossip-0.22.tar.gz&usg=AFQjCNGsht7VP5cFSfZdnOPL3mZRumIc5A

#639 July 21, 2008 at 6:16 pm in reply to: gossip protocol info
"> Details
Charles Pandian
Keymaster

Just try the following two links.

It will explain a very user friendly GUI based software called “ddd” – it will internaly use gdb for debugging.

http://mailman.isi.edu/pipermail/ns-users/2007-January/058498.html

and

http://pauseandreconsider.blogspot.com/2007/07/debugging-ns-2.html

But, I hope that, debugging ns2 code is not that much simple for a ‘beginner of ns2’.

#635 July 20, 2008 at 4:30 pm in reply to: about gdb
"> Details
Charles Pandian
Keymaster

First try the Marc Greis’s tutorial

(particularly section X. Creating Wired-cum-Wireless and MobileIP Simulations in ns)

It is available with ns2 (ns-allinone-2.29ns-2.29ns-tutorial)

You can see it online at the following link.

http://www.isi.edu/nsnam/ns/tutorial/

Hopefully you will solve your problem.

#631 July 18, 2008 at 1:49 pm in reply to: 802.11 simulation in NS2
"> Details
Charles Pandian
Keymaster

Of course, we have to use some old versions of ns2 for some kind of research. (ex : Still, ns2.1b is widely used for different kinds of research)

Generally, the ns2 architecture from version 2.26 is almost similar. So, we can port the code meant for one version of ns2 to another version with little effort.

But, if you are just studying video streaming only, then you may simulate it using video trace support available in the following link.

http://hpds.ee.ncku.edu.tw/~smallko/ns2/Evalvid_in_NS2.htm

You can even use the “video trace traffic” just directly without any modification on ns2. In that case you may simulate a model p2p network by using existing facilities available in ns2.

#628 July 18, 2008 at 2:40 am in reply to: How to make P2P Streaming or Overlay Network?
"> Details
Charles Pandian
Keymaster

GnutellaSim is designed for packet-level peer-to-peer system simulation

You may download it from the following link
http://www.cc.gatech.edu/computing/compass/gnutella/install.html

#626 July 17, 2008 at 5:37 pm in reply to: How to make P2P Streaming or Overlay Network?
"> Details
Charles Pandian
Keymaster


set traceFD	[open "test.tr" w]

$ns_ trace-all $traceFD



set namFD    [open "test.nam" w]

$ns_ namtrace-all-wireless $namFD $val(x) $val(y)



…..

……



close traceFD

close namFD

exec nam  test.nam

In the above code segment,

*The first line declares and opens file ‘test.tr” in write mode by using a file descriptor ‘tracefd’
*The second line will record all the trace events in that file using that file descriptor

*The third line declares and opens a file ‘test.nam” in write mode by using a file descriptor ‘namFD’
*The fourth line will record all the nam events in that file using that file descriptor

*The last three lines will close those files using those file descriptors and display the nam file.

Trace analysis will be made on the file “test.tr” for extracting information and that information will be used to display graphs using xgraph /gnuplot or any other tool.

Trace analysis can be done using any scripting language (any unix shell, tcl,perl,python,awk etc.,) or programming language(c, c++ java etc.,) of choice.
[/list]

#624 July 10, 2008 at 7:43 pm in reply to: reading the output file and trace file in ns2
"> Details
Charles Pandian
Keymaster

Writing a new module/code for queue management and adding it with existing ns implementation will require lot of skill and understanding about ns2 architecture.

As a beginner, one can implement their queue management by altering any one of the existing implementations of ns2 with somewhat lesser effort than writing entirely new modules.

#622 June 24, 2008 at 8:02 am in reply to: new queue management in ns2
"> Details
Charles Pandian
Keymaster

I hope that PHP is the best for web based application development for the following reason.

1.Since php is very native to linux/unix based platforms, almost all the related software like mysql, apache are free for the users. So the development cost related with software will be Zero.
2.Naturally all the linux/unix web related applications/software will provide better performance while comparing it with windows based software, the web application which will be developed using php will deliver better performance in terms of speed.
3.Even, the php version for windows also available on free of cost.

#621 June 24, 2008 at 7:44 am in reply to: PHP and C#
"> Details
Charles Pandian
Keymaster

The line

Scheduler::instance().schedule(target_, p,
0.01 * Random::uniform());

will Schedule the packet for transmission.

The packet will be sent to all the neighbors if it is marked (in the header part) as a broadcast packet.

#620 June 5, 2008 at 5:14 am in reply to: AODV Forward function
"> Details
Charles Pandian
Keymaster

You can implement it by adding suitable code at suitable place in any one of the following functions of aodv.cc

AODV::recvRequest(Packet *p) {

Or

AODV::forward(aodv_rt_entry *rt, Packet *p, double delay) {

#619 May 14, 2008 at 2:35 pm in reply to: Implement probabilistic broadcast
"> Details
Charles Pandian
Keymaster

As far as my understanding, the following lines Will not make the agent to behave as a routing agent. I hope still the default routing protocol DV will take care of routing.



set a0 [new Agent/MyAgentOtcl]

$ns attach-agent $n0 $a0

set a1 [new Agent/MyAgentOtcl]

$ns attach-agent $n1 $a1

So while you are sending something via your agent, then the default routing agent will only receive the packet and try to forward it. But while receiving this new packet type, it will not find the port/slot for dispatching the paket. So it can not forward this packet and will end up with the error

Classfier::no-slot{} default handler (tcl/lib/ns-lib.tcl)

It means that there should be a handler for your packet.

Generally while writing a new routing protocol (MANET) , there may be some necessary modifications required in some files like tcl/lib/ns-lib.tcl and some files under /common directory.

See the following link for installing a MANET routing agent called NOAH

http://icapeople.epfl.ch/widmer/uwb/ns-2/noah/

Kindly post your reply if I am wrong or while you find a solution.
[/code]

#618 April 30, 2008 at 8:12 am in reply to: creating routing agent
"> Details
Charles Pandian
Keymaster

This mean that you not yet successfully patched ns2 for MAODV. Repeat the installation/Patching procedure of MAODV again and then try the tcl script.

#612 April 29, 2008 at 9:35 pm in reply to: MAODV over cygwin??
"> Details
Charles Pandian
Keymaster

If you want to implement it at C++ level ns2 code, then you may use array or structure in the class file corresponding to the node

For example in “node.cc” and “node.h” (if database only). Or you may implement it in the routing agent code itself. (ex : aodv.cc and aodv.h). But, this will require recompilation of ns2.

Even you can do the same thing inside the tcl part of simulation script also. This will require any recompilation of ns2.

#617 April 29, 2008 at 9:22 pm in reply to: node as a database
"> Details
Charles Pandian
Keymaster

You not at all mentioned your question in your post.

#615 April 29, 2008 at 9:14 pm in reply to: analysis of trace file for MANET
"> Details
Charles Pandian
Keymaster

You need not do everything in god.cc itself.

You may handle the values inside your new agent code (if any) or pass them to a tcl variable as follows. Here AddNeighborToList will add the neighbors to a tcl list. After that You can use that variable from the simulation script itself.


The c++ part





void MyAgent::FillNeighborList()

{

Tcl& tcl = Tcl::instance();



int TotalNodes=God::instance()->nodes();



for(int i=0;i
{

if(God::instance()->IsNeighbor(i , here_.addr_))

{

tcl.evalf("%s AddNeighborToList %d ",name(), i);

}

}

}


the tcl part



Agent/MyAgent  instproc AddNeighborToList {nn} {

$self instvar NeighborList_

lappend NeighborList_ $nn

}
#609 April 21, 2008 at 3:49 pm in reply to: Asking about number of neighbors
"> Details
Charles Pandian
Keymaster

I hope that, you need not implement your won version of flooding. Because, there is already there are tcl level examples for flooding under
ns-x.xxtclex directory.

There is also another flooding based routing agent available under
ns-x.xxdiffusion

#607 April 11, 2008 at 3:10 pm in reply to: How to implement flooding for wired nodes (IP_BROADCAST usag
"> Details
Charles Pandian
Keymaster

I hope that the characteristics of WBAN mentioned in

https://www.nari.ee.ethz.ch/wireless/research/projects_01/ban.html

can be simulated using existing Sensor Network Support of ns2.

#606 March 19, 2008 at 2:51 am in reply to: ns2 code for WBAN
"> Details
Charles Pandian
Keymaster

An agent code implemented in CC file can be integrated into ns2 and can be inherited or used from a tcl file.

The ns2 interpreter can “run” only tcl scripts not CC files.

#605 March 12, 2008 at 3:17 pm in reply to: Help me ….. Urgent……….
"> Details
Charles Pandian
Keymaster

Thank you for this information.
It will be very useful for research students.

#604 March 11, 2008 at 6:49 pm in reply to: A new simulator
"> Details
Charles Pandian
Keymaster

Your question is not clear. If you read your question from another person’s point of view, you can certainly realize that.

1. If you want to print anything from a c++ file, then you can use just use “printf”
2. If you want to print anything from a tcl script, then you can use just use “puts”

3. If you want to pass arguments to a tcl script, then Tcl scripts have access to the following predefined variables. Just learn to use it inside the script.
$argc – number items of arguments passed to a script.
$argv – list of the arguments.

4. If you want to pass a value from c++ to tcl then you can use bind command in c++ file as follows.
bind (“myvar_”, &myvar_);

You can use the myvar inside your tcl script to pass a value to c++ code or access a c++ value from tcl script. (But it is not simple so Learn more about this to do this)

Anyhow, still, I don’t understand your question.

#603 February 28, 2008 at 10:21 am in reply to: Reading the arguments from Tcl file in a C++ file
"> Details
Charles Pandian
Keymaster

There are some mobile IP extensions available for ns2. (like Hawaii)

See the following links
http://typo3.cs.uni-paderborn.de/fileadmin/Informatik/AG-Karl/Pubs/WNET2004.pdf

http://core.it.uu.se/core/index.php/AODV-UU_and_Mobile_IP_for_ns-2

http://www.isi.edu/nsnam/dist/gprs-072501-2.1b7a.tar.gz

But most of these exertions will work only on old versions of ns2.

You may try to do your simulation by improving any one of this.

#601 February 27, 2008 at 9:41 am in reply to: regarding handoff
"> Details
Charles Pandian
Keymaster

You can use any text editor to edit .bash_profile.

#vi .bash_profile

or

#gedit .bash_profile – if gedit is in your system

The script will get executed automatically while login again in the system.

#600 February 26, 2008 at 11:14 am in reply to: How to edit bash_profile
"> Details
Charles Pandian
Keymaster

If there will not be any readymade solution available for this,

After some effort, I hope that it can be done by adding suitable code in aodv.cc in suitable places at the following functions.


AODV::sendRequest (…)
and
AODV::recvAODV(…)

Certainly it will not be a simple task

#599 February 26, 2008 at 7:05 am in reply to: how to know route request and route respond time
"> Details
Charles Pandian
Keymaster

First Try to increase the bandwidth, queue size and increase the transmission range of the nodes or decrease the distance between nodes. Hopefully it will solve the problem.

#596 February 19, 2008 at 4:30 am in reply to: multichannel
"> Details
Charles Pandian
Keymaster

If you are doing any change in any of the files of ns2, then don’t compile it separately. Instead of it, just change to the ns-allinone-2.xxns-2.xx directory and run make. It will use the Makefile and compile the whole project and create a new binary version of ns2 in the same directory. After that, you can install the new version of ns2 into appropriate path using “make install” command. Some times you have to run the script “configure” before make. Run the commands as follows.

#./configure (dot slash configure)
– the “configure” step is required only once and may not necessary if you are working on a previous installation of ns2.
#make
#make install

#592 February 18, 2008 at 5:54 pm in reply to: Error Compile aodv.cc
"> Details
Charles Pandian
Keymaster

It seems that the following SIP path for ns2 contains some implementation of DNS.

http://www-x.antd.nist.gov/proj/iptel/ns2-sip-simulation/ns-2.1b9a.sip.patch

#587 February 18, 2008 at 4:27 am in reply to: Detecting DoS attacks against DNS
"> Details
Charles Pandian
Keymaster

If xgraph has been installed properly on a linux/unix/Cygwin system, then you can use the “man” command to see the manual page of xgraph.

That is, the command

#man xgraph

will show several pages of help for using xgraph.

You will see all the parameters there.

#586 February 17, 2008 at 4:53 pm in reply to: parameters 4 xgraph
"> Details
Charles Pandian
Keymaster

Check the following links

http://www.cse.msu.edu/~wangbo1/ns2/nshowto8.html

http://my.opera.com/HenryFD/blog/show.dml/155516

http://blog.pixnet.net/zayrowz/post/10527361

#584 February 15, 2008 at 4:47 pm in reply to: multichannel
"> Details
Charles Pandian
Keymaster

See

http://projectguideline.com/ns2.htm Page.

You will find links to some standard tutorials.

#582 February 8, 2008 at 1:34 am in reply to: help in ns2
"> Details
Charles Pandian
Keymaster

If it is possible, upload a simple version of a tcl script demonstrating this scenario and what output/error you are getting while running the script. (Please Mention the version of your ns2)

#574 February 3, 2008 at 3:36 am in reply to: ns Multicast in wired+wireless scenario
"> Details
Charles Pandian
Keymaster

It is some what tough and is called as trace analysis. For doing this we have to have good knowledge about ns2 trace file formats.

Filter the necessary events from the trace file using scripting languages like tcl, perl, Python and awk or shell scripts or any other languages like C, C++, Java and prepare two column data file (one is for x and another is for y). now call xgraph with the data file as follows.
# xgraph mydata.txt

here the mydata.txt will contain the two column data.

#573 January 31, 2008 at 1:46 pm in reply to: tracefile interpretation
"> Details
Charles Pandian
Keymaster

Have you tried this one?

http://nile.wpi.edu/NS/new_app_agent.html

#580 January 31, 2008 at 1:37 pm in reply to: ns2 How to create new agents???
"> Details
Charles Pandian
Keymaster

Generally, the new agent will be attached with a nodes.

If the receive function of a node or the agent attached with that node is receiving something, then at that point it can refer its variable only, That is , if the lines

$zigbee set DestinoPaqID 79
$zigbee set OrigenPaqID 2

refer that particular agent of the node, then the value at that post will be 79 and 2.

Other wise if you are setting these the values for an agent of a different node and referring those values from some other node, then it will show the default value only.

#572 January 27, 2008 at 8:03 am in reply to: bind variables
"> Details
Charles Pandian
Keymaster

You did not mention how you are changing those values in your simulation.

Generally if the node is moving out from the topographical area, ns will raise error.

#571 January 22, 2008 at 2:32 pm in reply to: Segmentation fault in ns
"> Details
Charles Pandian
Keymaster

Hello Dhaya

Your question and the code were more irrelevant.

You are telling that you are Testing the script for routing agent TORA.

But your script is using SWARM Intelligence related routing agent.

“set val(rp) SWARM ;# routing protocol”

#570 January 22, 2008 at 7:58 am in reply to: help regarding TORA protocol
"> Details
Charles Pandian
Keymaster

Generally, in ns2, this kind of task can not done only with TCL scripts. You have to write the required code at specific areas of existing agents or you have to write your own agent to do the same.

For example, the following code segment is handling ttl.





hdr_ip *iph = HDR_IP(p);

hdr_cmn *ch = HDR_CMN(p);

int src = Address::instance().get_nodeaddr(iph->saddr());

int dst = ch->next_hop();



/*

*  Must be a packet I'm originating...

*/

if(src == myaddr_ && ch->num_forwards() == 0) {

/*

* Add the IP Header

*/

ch->size() += IP_HDR_LEN;

iph->ttl_ = IP_DEF_TTL;

}

/*

Learn to write or manipulate agent code in ns2.
[/code]

#566 January 11, 2008 at 7:35 pm in reply to: how to read header information?
"> Details
Charles Pandian
Keymaster

Have you applied the Patch for FC5 before starting the installation?

it is available at the following link

http://www.ececs.uc.edu/~cdmc/ucbt/src/ns2292-gcc410.patch

for more info refer

http://nsnam.isi.edu/nsnam/index.php/Ns_Users_FAQ

#554 January 10, 2008 at 4:41 am in reply to: Get tcl error during installing ns2.29 on Fedora and ubuntu
"> Details
Charles Pandian
Keymaster

Check the following link for a hint on Manual Routing.

http://www.isi.edu/nsnam/ns/ns-largesim.html

#547 January 8, 2008 at 3:46 pm in reply to: modify manual routing
"> Details
Charles Pandian
Keymaster

If you mean a ‘TCL Script” to do this kind of election, then my answer will be NO. Probably you may not find such a script on Internet since these functionality were generally implemented with complex C++ Code.

#552 January 8, 2008 at 3:22 pm in reply to: script of root node election..!!
"> Details
Charles Pandian
Keymaster

I hope that you don’t need to implement a timer.

Refer the code of routing protocols such as DSR. There you may find the usage of a timer.

For example, in DSR, you will find this timer code in DSRAgent.h





class SendBufferTimer : public TimerHandler {

public:

SendBufferTimer(DSRAgent *a) : TimerHandler() { a_ = a;}

void expire(Event *e);

protected:

DSRAgent *a_;

};
#553 January 8, 2008 at 3:16 pm in reply to: timer implementation
"> Details
Charles Pandian
Keymaster

Before starting with your own code, first try to implement existing QoS additions (if any) on a suitable version of ns2. It will give you better understanding of the architecture of ns2.

Instead of writing your own code from single line, you can alter a suitable QoS extension as per your needs.

#551 January 7, 2008 at 4:58 am in reply to: QoS in Manets
"> Details
Charles Pandian
Keymaster

In the official Ns2 manual Chapter 7 (Queue Management and Packet Scheduling) is explaining implementation details of the queue in ns2.

Download the document from the following link.

http://www.isi.edu/nsnam/ns/doc/ns_doc.pdf

#550 January 7, 2008 at 4:51 am in reply to: need clarification in droptail..
"> Details
Charles Pandian
Keymaster

Sorry. Your Question is not clear.

what’s there in ‘ droptail’ in the folder ‘queue ‘ of ns2 package?

As the name signifies that there will be code related with a queue management scheme named droptail.

Please ask some thing more specific.

#548 January 5, 2008 at 2:24 pm in reply to: need clarification in droptail..
"> Details
Charles Pandian
Keymaster

Manual routing can be done using a routing agent called NOAH. You can download it and integrate it with ns2.

Matlab can be used from ns2. You can learn it from following link

http://www.cse.msu.edu/~wangbo1/nshowto3.html

But, interacting with matlab from network layer is not advisable because, that kind of interaction will not be suitable for packet level processing and decision making required at network layer.

Charles Pandian,
IGS

#545 January 4, 2008 at 5:49 am in reply to: modify manual routing
"> Details
Charles Pandian
Keymaster

Have you tried the following link?

http://www.ece.rochester.edu/research/wcng/code/index.htm

#544 January 3, 2008 at 5:24 pm in reply to: LEACH ns-2 Code please!!!
"> Details
Charles Pandian
Keymaster

Sine DCHP is a application is a simple service of Tcp/IP protocol, it can be done with little effort while comparing with the effort needed for other things such as a routing protocol.

It can be implemented using an agent called “Agent/MessagePassing” you can see an example script in ns2 which will use that agent..

In this forum also, I have explained about that agent in some other thread.

#543 December 21, 2007 at 4:01 pm in reply to: Creating DNS agent in top of UDP
"> Details
Charles Pandian
Keymaster

If you are adding “ex-linkage.o” then you have to refer the full path.

That is if you are keeping the files ex-linkage.cc and ex-linkage.h under the directory
ns-allinone-2.30ns-2.30test

then you have to mention it as follows

test/ex-linkage.o

#542 December 19, 2007 at 9:57 am in reply to: make failed
"> Details
Charles Pandian
Keymaster

Open the Makefile in vi editor (not other editor).
See the line you have just inserted(object.o in Makefile (OBJ_CC)).
Probably you will see ^M character at the end of the file.
Remove that character and compile again.

If you do not know vi editor, then open the file in any gui editor and position the cursor at the end of the line and delete all the characters(It will delete the invisible ^M). (It will bring the next line to the current line. So just press enter to bring the next line to its original position). Repeat the same procedure for all the lines where you have inserted a new line.

The Procedure may seems to be strange but it solved my problem in several occasions. Probably it will solve your problem also.

Reason (as far as I believe): I faced the problem while doing cut and past from a dos file to unix file.

Some times make will be successful even if we are having ^M at some other places. I don’t know why. If some one is familiure with the problem, then kindly discuss your experience here.

#541 December 19, 2007 at 3:13 am in reply to: make failed
"> Details
Charles Pandian
Keymaster

What about the total size(bytes) of the recieved packets at the point of interest in both case?

#540 December 18, 2007 at 1:23 pm in reply to: Abour CBR in wired-cum-wireless
"> Details
Charles Pandian
Keymaster

Have you seen the following link for your answer?,

http://www.isi.edu/nsnam/archive/ns-users/webarch/1999/msg02666.html

#538 December 18, 2007 at 1:08 pm in reply to: Heap Scheduler
"> Details
Charles Pandian
Keymaster

Try the following. You may solve your problem.

http://www.lst.inf.ethz.ch/research/publications/publications/MSWiM_2003/MSWiM_2003.pdf

http://www.cs.binghamton.edu/~kliu/research/ns2code/index.html

Chrles Pandian.

#536 December 13, 2007 at 2:35 am in reply to: MIP headache in ns-2
"> Details
Charles Pandian
Keymaster

It is .bashrc not ./bashrc.
It will be in your home directory.

To go to your home directory, just issue the command ‘cd’ without any parameters.

Instead of using ./bashrc use can also .bash_profile which will be in your home directory.

In unix anything starting with a . ( period) will be considered as a hidden file. so just issue the command vi .bash_profile in your home directory. It will open the file. If it is not there, just create one.

Charles Pandian.

#535 November 25, 2007 at 5:45 pm in reply to: NS-2 on Linux installation problem
"> Details
Charles Pandian
Keymaster

in the file
ns-allinone-2.28ns-2.28dsrdsragent.cc,

See all the functions which are calling a function
sendOutPacketWithRoute(…).

You may find your answer.

#533 November 7, 2007 at 6:16 pm in reply to: Appending a node address in DSR protocol
"> Details
Charles Pandian
Keymaster

Two days back, I receive the following message from paypal


“Now you can add your Indian bank account to your
PayPal account, so you can withdraw your money
directly to your bank. It saves you time and gives
you faster access to your PayPal funds!

Plus, there is no charge to withdrawal amounts
over INR 7000*. So, go ahead and add your bank account today! “

I hope that it will be legal since it seems that the Indian banks are accepting that transaction.

Some one please confirm it.

#515 November 5, 2007 at 7:04 am in reply to: Is PayPal Legal in India?
"> Details
Charles Pandian
Keymaster

Set the environment variables in .bash_profile in the home directory of the user(the directory in which you are in after the login process). It is a hidden file so to see the file use

#ls –a

edit it using vi editor

#vi .bash_profile

add the necessary environment variables
(refer the last pages of install script of ns2 – the file ns-allinone-2.28ns-2.28 install)

If you do it properly, then xgraph and other executables of ns2 can be accessed from any diroctory.

To show a graph for two column data file test.dat, run the command as follows.

#xgraph test.dat

#529 November 5, 2007 at 6:56 am in reply to: need to know tcl script programming
"> Details
Charles Pandian
Keymaster

If you mean ns2 related tcl commands, then start with the Famous ns2 Tutorial by Marc Greis

ns-allinone-2.28ns-2.28ns-tutorial

You may find some examples under ns2 directory.

ns-allinone-2.28ns-2.28ns-tutorialexamples
ns-allinone-2.28ns-2.28tclex

You can learn tcl from tutorials available in internet.

Since tcl is a interpreted programming language, actually there is no compilation step. Just you can directly run your tcl script (here the ns2 simulation script) by using the ns2 interpreter as follows.

$ns program.tcl

#527 October 1, 2007 at 1:39 pm in reply to: need to know tcl script programming
"> Details
Charles Pandian
Keymaster

Once I asked the same questing at one of my old student who is having a habit of switching jobs very frequently. He said, “sir, in interviews, for a complex technical question, if one is able to say the answer – ‘I don’t know’ – in an indirect way with better English and style then most of the time, it will be a good and satisfactory answer”.

💡 If the question is ‘The Role of Communication Skills in Getting an IT Job’ then, option ‘B’ only is the answer.

💡 But, I believe that option ‘B’ is just like a ‘key’ to open a workshop door. In side the workshop we can not do the work only with the key. Inside the workshop we have to be equipped with some additional ‘tools’. So option ‘A’ is also needed to survive after passing the door.

#523 September 2, 2007 at 5:57 pm in reply to: PROGRAMMING SKILLS OR COMMUNICATION SKILLS ?
"> Details
Charles Pandian
Keymaster

The following code segment shows the usage of Message Passing Agent.

You can refer the whole example script ‘flooding.tcl’ under the directory ns-allinone-x.xxns-x.xxtclex





# subclass Agent/MessagePassing to make it do flooding



Class Agent/MessagePassing/Flooding -superclass Agent/MessagePassing



Agent/MessagePassing/Flooding instproc send_message {size msgid msg} {

$self instvar messages_seen node_

global ns MESSAGE_PORT



$ns trace-annotate "Node [$node_ node-addr] is sending {$msgid:$msg}"



lappend messages_seen $msgid

$self send_to_neighbors -1 $MESSAGE_PORT $size "$msgid:$msg"

}



Agent/MessagePassing/Flooding instproc send_to_neighbors {skip port size data} {

$self instvar node_



foreach x [$node_ neighbors] {

set addr [$x set address_]

if {$addr != $skip} {

$self sendto $size $data $addr $port

}

}

}



Agent/MessagePassing/Flooding instproc recv {source sport size data} {

$self instvar messages_seen node_

global ns



# extract message ID from message

set message_id [lindex [split $data ":"] 0]



if {[lsearch $messages_seen $message_id] == -1} {

lappend messages_seen $message_id

$ns trace-annotate "Node [$node_ node-addr] received {$data}"

$self send_to_neighbors $source $sport $size $data

} else {

$ns trace-annotate "Node [$node_ node-addr] received redundant copy of message #$message_id"

}

}
#521 August 19, 2007 at 6:39 pm in reply to: Send real stuff in NS2
"> Details
Charles Pandian
Keymaster

Refer the function God::IsNeighbor(int i, int j) in mobilegod.cc you will have some idea about neighbor discovery in MANET.

Charles Pandian.

#513 August 2, 2007 at 7:01 am in reply to: How to acess a set of neighbors of a node in AODV?
"> Details
Charles Pandian
Keymaster

Generally ns2 is only tested on some specific version of Unix variants (like Red Hat Linux). But after particular release of ns2 version, research people around the world may give a patch file (a script 💡 which will automatically fixes ns2 for installing it in a different variant of Unix/Linux 😀 ).

Patching a ns2 version for particular Linux variant is usually will not be a simple task 😕 . Further even after successful compilation, we may face different kinds of problems while running the simulator.

➡ If you are a beginner and having short time to complete your assignment or project, then try ns2 on a standard tested version of linux (for example, ns2.28 will install successfully on Red Hat Linux 9 without the necessity of patching. Simularly, some new version of ns2-s will work on some new linux variants without the necessity of patching – refer the ns2 home page)

➡ Finally, I don’t see any readymade patch available for centos5 for any versions of ns2.

If you are very specific about ns2.28 on centos5 then confirm it again in this thread.

#511 July 22, 2007 at 4:38 am in reply to: NS installing problems
"> Details
Charles Pandian
Keymaster

Put

$ns_ run

as the last line of the script. Because you have to run the simulation then only it will create some output.. Is isn’t it?. Further, in your code, you did not do anything for nam visualization – you don’t want nam output?.

These are very basic mistakes.

So first take some time to read ns2 Tutorial by Marc Greis ( http://www.isi.edu/nsnam/ns/tutorial/ ) and understand the basic constituents of ns2 simulation script. Try the samples provided in that tutorial.

#510 July 16, 2007 at 8:15 am in reply to: Error when simulating the generated files from “cbrgen.
"> Details
Charles Pandian
Keymaster

I downloaded the pdf document from the link.
Is it available as a patch to any version of ns2?

How is it differing from the IIT’s ‘enhanced network simulator’ – TENS.?

#503 July 16, 2007 at 4:06 am in reply to: Simulation times and traffic stream label
"> Details
Charles Pandian
Keymaster

Which version of the multi-channel extension you are using?

#501 July 14, 2007 at 6:10 pm in reply to: Simulation times and traffic stream label
"> Details
Charles Pandian
Keymaster

Justt cut the following code segment and past it on your previous code. It will work. I have corrected and tested it on ns2.31.

Can you guess what is the error? I will tell you latter.

set god_ [create-god $val(nn)]

$ns_ node-config -adhocRouting $val(rp) -llType $val(ll) -macType $val(mac) -ifqType $val(ifq) -ifqLen $val(ifqlen) -antType $val(ant) -propType $val(prop) -phyType $val(netif) -channelType $val(chan) -topoInstance $topo -agentTrace ON -routerTrace ON -macTrace OFF -movementTrace OFF

#508 July 13, 2007 at 7:03 pm in reply to: Error when simulating the generated files from “cbrgen.
"> Details
Charles Pandian
Keymaster

If the transmission power is very low, then the nodes will not receive
anything and hence could not involve in any communication. the following
table can be used for calculating the distance with respect to the Pt_ , and
it is a rough estimate.

Let us assume that a receiving node is set at this following parameters.

Phy/WirelessPhy set RXThresh_ 630e-15
Phy/WirelessPhy set CSThresh_ 630e-15

This means, that node only consider a carrier signal which is having energy
630e-15 or higher and forward the received frame if the signal energy is
630e-15 or higher.

By increasing transmission power, the signal may reach a node but it will be
ignored if the signal strength is less than RXThresh_ or CSThresh_ at
receiving end.

#505 July 13, 2007 at 1:26 pm in reply to: Need help with the cell size in NS2
"> Details
Charles Pandian
Keymaster

While runing your script on ns2.28, 2.29 and 2.31, it is showing following error. ( I will try it on ns2.30 latter).

num_nodes is set 2

(_o3 cmd line 1)
invoked from within
“_o3 cmd change-numifs 2”
invoked from within
“catch “$self cmd $args” ret”
invoked from within
“if [catch “$self cmd $args” ret] {
set cls [$self info class]
global errorInfo
set savedInfo $errorInfo
error “error when calling class $cls: $args” $…”
(procedure “_o3” line 2)
(SplitObject unknown line 2)
invoked from within
“$ns_ change-numifs $val(ni)”
(file “test2.tcl” line 41)

You can set the label of a node as we set the color previously.

while running a simulation, if there is nothing for recording (ie, no traffic due to some reason), then the simulation will not show anything beyond that even if you are running it for some more time.

#499 July 13, 2007 at 12:49 pm in reply to: Simulation times and traffic stream label
"> Details
Charles Pandian
Keymaster

Explain the problem Briefly and specify what kind of ns2 simulation you are requiring. I will explore the possibilities of implementation.

#497 July 13, 2007 at 10:11 am in reply to: Scheduling Algorithm in Wireless communication
"> Details
Charles Pandian
Keymaster

what did you mean by partial compilation of ns2 trace related files?
Ans : if you are trying to add some new agent or protocol, if there is some change in trace support, then it should al so be done properly. In some cases, the simulation will run without the new trace support but, the output trace file will not reflect some events.

Should its value be one of the node id which is from 0 to 7
Ans : No. according to manual it is not node id.

I use NOAH routing protocol for wimax configuration. Is this the reason?
Ans : While adding NOAH routing protocol there is no additions for trace support. It means the packets of NOAH will use one of the existing trace supports. So probably it will not be a problem.

But I can’t understand why the trace format for some wireless link are shorter than the standard one.
Ans : probably it is due to the mixed scenario. Wired node trace is little bit different from wireless node trace.

#492 July 4, 2007 at 1:05 pm in reply to: Need help with the wireless tracefile.
"> Details
Charles Pandian
Keymaster

With phrase “periodic route interval” ❓ – you mean the route table update time, beacon broadcast interval etc.,?. If it is the case, you may find lot of timers in c code of every routing protocol. If there is any TCL command interface provided to handle those timers then we can use them from TCL script 😀 . Otherwise if we change one such interval or any other “parameter”, then it will require re-compilation 🙄 of ns2.

#489 June 26, 2007 at 7:36 am in reply to: regarding the constants used in the protocols
"> Details
Viewing 264 posts - 1 through 264 (of 264 total)
WhatsApp Discuss Through WhatsApp