Hi,
I got it. Page 127 of ns-doc.pdf tells us how to do it. But it's
really hard to find it. I thought it is covered in "Chapter 16 Mobile
Networking in ns" or "Chapter 6 Links: Simple Links"
Another question,
How long does it take to transmit a packet? I found this snippet in
ns/tcl/ex/wpan:
set expl($src) [new Application/Traffic/Exponential]
eval \$expl($src) set packetSize_ 70
eval \$expl($src) set burst_time_ 0
eval \$expl($src) set idle_time_ [expr
$interval*1000.0-70.0*8/250]ms ;# idle_time + pkt_tx_time = interval
eval \$expl($src) set rate_ 250k
the packet size is set to 70 and pkt_tx_time is calculated as
70.0*8/250, but according to the wpan_demo2.tr file the packet(MAC)
size is 97 bytes, then should we calculate the pkt_tx_time as
97.0*8/250 ? Or is it unnecessary to consider the mac header?
On 4/3/06, Ahmad Khayyat <[EMAIL PROTECTED]> wrote:
> Hi Bruce,
>
> Read chapeter 13 in ns-manual.. I think you are looking for Error Model
> instead of Loss Model.
> The commands you are using are for wired networks only, if I am not
> mistaken.
>
>
> Bruce Who wrote:
> hi, Ahmad:
>
> ======= 2006-04-03 02:05:24 Ahmad Khayyat wrote: =======
>
>
>
> Hi Bruce,
>
> If you enable AGT and RTR tracing (in node-config command in the tcl
> file), you will see that when a tcp packet is sent out of the RTR layer,
> its size is larger than its size when sent by the AGT layer. This
> reflects the fact that the RTR layer adds some headers that contribute
> to the packet size. The same thing happens when the packet is sent by
> the MAC layer; it gets a little bigger due to MAC headers.
>
> For the inconsistent packet size in the MAC layer between send time and
> recv time, my guess is that the packet size is measured in the receive
> event after removing the MAC headers. You will see that the received
> packet size in the MAC is equal to the packet size sent by RTR, which is
> the size before adding the MAC headers. Having that said, my guess is
> that the packet size reported in the send event is the correct size sent
> out of the MAC layer. In that case, the packet size in the receive event
> is the size that the MAC layer reports to higher layers.
>
> I agree with you. I find that all beacon/ack packets are exactly 7 bytes.
> I guess this is a bug of ns2?
>
> And I still cannot apply lossmodel to wpan. I modified
> ns-2.29\tcl\ex\wpan\wpan_demo2.tcl this way:
>
> =============================================================
> proc seterrmodel { src dst } {
> global ns_ val node_
> set errmodel [new ErrorModel]
> $errmodel set rate_ $val(per)
> $errmodel set unit_ $val(errUnit)
> eval $ns_ lossmodel $errmodel \$node_($src) \$node_($dst)
> }
>
> ...
>
> seterrmodel 0 1
> =============================================================
>
> I added above code into wpan_demo2.tcl, and got the following traceback:
>
> $ ns wpan_demo2.tcl
> num_nodes is set 7
> INITIALIZE THE LIST xListHead
>
> Traffic: ftp
> Acknowledgement for data: on
> invalid command name ""
> while executing
> "$link errormodule $lossobj"
> (procedure "_o3" line 3)
> (Simulator lossmodel line 3)
> invoked from within
> "_o3 lossmodel _o158 $node_(0) $node_(1)"
> ("eval" body line 1)
> invoked from within
> "eval $ns_ lossmodel $errmodel \$node_($src) \$node_($dst)"
> (procedure "seterrmodel" line 7)
> invoked from within
> "seterrmodel 0 1"
> invoked from within
> "if { "$val(traffic)" == "ftp" } {
> puts "\nTraffic: ftp"
> #Mac/802_15_4 wpanCmd ack4data off
> puts [format "Acknowledgement for data: %s" [Mac/8..."
> (file "wpan_demo2.tcl" line 231)
>
> Could anyone give me any clue?
>
>
>
> Bruce Who wrote:
>
>
>
> hi, Ahmad:
>
> ======= 2006-03-31 23:40:43 Ahmad Khayyat wrote: =======
>
>
>
>
>
> 2) assoPermit means whether the node should permit other nodes to
> associate with it so that it becomes their coordinator. Look at the
> 802.15.4-2003 standard for details.
>
> 5) This is not how LossModel is used. Consult ns-manual for the correct
> usage. It still applies for wpan.
>
>
>
> Thanks, I'll look into the document.
>
> And I have another question. This is a snippet from wpan_demo2.tr, I find
> that all tcp packets sent out is always 7 bytes bigger than the received
> packets. Why do they not have the same size? Then which one(97 or 90) is the
> real size of the tcp packets?
>
> s 11.928192000 _0_ MAC --- 255 tcp 97 [0 5 0 800] ------- [0:2 5:0 30 5]
> [133 0] 0 0
> s 11.931712033 _5_ MAC --- 255 ACK 5 [0 0 5 0]
> r 11.932064067 _0_ MAC --- 255 ACK 5 [0 0 5 0]
> r 11.932704033 _5_ MAC --- 255 tcp 90 [0 5 0 800] ------- [0:2 5:0 30 5]
> [133 0] 1 0
>
>
>
>
>
> Bruce Who wrote:
>
>
>
>
>
> ------------------------------------------------------------------------
>
> Hi, all:
>
> Recently I begin to learn how to use ns2 to run wpan simulations, and
> stumbled
> by some problems, hope someone could help me out.
>
> 1)I found there are some commands like this:
> eval $ns_ attach-agent \$node_($src) \$udp_($src)
>
> why don't we just use this command?
>
> $ns_ attach-agent $node_($src) $udp_($src)
>
> 2) for this command
> $node sscs startCTDevice <isFFD = 1> <assoPermit = 1> <txBeacon = 0>
> <beaconOrder = 3> <SuperframeOrder = 3>
>
> what does assoPermit mean?
>
> 3) I ran wpan_demo2.tcl and found that all extended addresses of nodes are
> [0]. Is it necessary to set the address to different values by ourselves?
>
> 4) I ran "nam wpan_demo2.nam" and found that data stream is only seem
> between
> PANCoor(node0) and node1. But we have such statements in the wpan_demo2.tcl
> script:
>
> ftptraffic 0 1 $appTime1
> ftptraffic 0 3 $appTime3
> ftptraffic 0 5 $appTime5
>
> So why could no communications be seen between node0 and node3/node5?
>
> 5) And how to set lossmodel for nodes in wpan? I tried following code, but
> it
> doesn't work:
>
> set loss_module [new ErrorModel]
> $loss_module set rate_ 0.01
> $loss_module unit pkt
> $loss_module ranvar [new RandomVariable/Uniform]
> $loss_module drop-target [new Agent/Null]
> eval $ns lossmodel $loss_module \$node_($src) \$node_($dst)
>
> Any help would be appreciated! Thanks in advance!
>
>
>
> Best regards,
>
> Bruce Who
>
>
>
>
>
>
> = = = = = = = = = = = = = = = = = = = =
>
> Best regards,
>
> Bruce Who
> 2006-04-02
>
>
>
>
> = = = = = = = = = = = = = = = = = = = =
>
> Best regards,
>
> Bruce Who
> 2006-04-03
>
>
>
