[ns] Rate of errors with an error model in wireless simulatio
n.
Maneesh Tewari
tmaneesh@cedt.iisc.ernet.in
Mon Apr 22 05:18:20 2002
hello sumitha,
the error module specified in the user's tcl file determines whether a
packet is contaminated by running some random function. (Check with corrupt
procedure in errmodel.cc) If the result of the random function is less than
the parameter rate_, the packet's error field is set to 1. Then the
mac802.11 module of the receiver discards the packet with the error field
set to 1. (Check with discard procedure in mac-802_11.cc).
its all the magic of random no.s the rate_ parameter has no direct relation
with increasing or decreasing the error on the wireless link, it may be
wrong but this is what i've experienced.
bye.
---maneesh
-----Original Message-----
From: Sumitha Bhandarkar [mailto:sumitha@ee.tamu.edu]
Sent: Sunday, April 21, 2002 8:53 AM
To: ns-users@ISI.EDU
Subject: [ns] Rate of errors with an error model in wireless simulation.
I have a very simple topology with just two mobile nodes, with one of the
nodes configured to have a uniform error model. The rate_ parameter is set
to 0.2, but the number of drops observed on the link is only about
0.006. When I increase the rate_ to 0.4, the observed drop rate
increases to 0.04, and When the rate_ is increased to 0.6, the approximate
droprate observed on the link is about 0.38. Has anyone else had this
disparity ? Isnt the rate_ parameter used to set the rate of packets
dropped on the incoming/outgoing link associated with the node ? Does
anyone know why the observed rate is so much lesser than the rate_ set
and how to resolve it ?
The tcl script I used for the error model is as follows -
ErrorModule set debug_ false
ErrorModel set enable_ 1
ErrorModel set markecn_ false
proc UniformErrorProc {} {
global opt
set errObj [new ErrorModel]
$errObj set rate_ 0.2
$errObj unit packet
return $errObj
}
$ns_ node-config -errProc $val(err)
Any input on this will be appreciated.
Thank You,
Sumitha.