[ns] TwoStateMarkov

[CWen56 <n2688420@dec4000.cc.ncku.edu.tw>]

Dear everybody:
    I have a question about definition of Two state markov error model.
I found a article in maillist as the following. But while i check
ns-errmodel.tcl, i find something strange : It should be
/Error/TwoStateMarkov not /Error/MultiState/TwoStateMarkov. But most
difficult for me is that how to use TwoStateMarkov ??
I can't realize what parameters should be used. And according to the
following article, what does that " rate " mean for??  especially it
sets rate [list 100000000 1] . It really confuses me.
Did any one ever use /TwoStateMarkov? Hope someone could give me a
hint!!

best regard!



************************************************
"Two state markov error model"

Javier Gomez-Castellanos (javierg@comet.columbia.edu)
Mon, 27 Apr 1998 20:04:15 -0400 (EDT)

I have a general question regarding the Two state markov error model:

if i define a model like this:

set trans [list 0.001 0.999]
set rate [list 100000000 1]
set unit byte
set em [new ErrorModel/MultiState/TwoStateMarkov $rate $trans $unit]

The error rates in each state look ok but, i could not find any logic
regarding the transition rates. In this particular example I expected to

have 1000 bytes correct/1 byte error/1000 bytes correct/one byte
error...