Re: [ns] Reg bandwidth utilization!!

[Ganesh <ggopalak@uiuc.edu>]

Hi ,

I got two questions...

1) For this to work, dont we have to declare bytes_ in tcp-sink.cc and
tcp.sink.h ?? And increment it accordingly ?? Or is it a different bytes_
that you are referring to ? and this gives me the effective throughput /
goodput seen  by the sink , right ?

2) How do we measure end-end delay using this method ? Which variable gives
me this info ?

Thanks.

- Ganesh

"T. Chitti Babu" wrote:

> ANOTHER OPTION WOULD BE TO...
>
> set f0 [open src0.tr w]
>
> proc finish {} {
>         global ns nf f0
>         $ns flush-trace
>         close $f0
>         exec xgraph -m src0.tr  -geometry 800x400 &
>         exit 0
> }
>
> #Define a procedure which periodically records the bandwidth received by
> the #traffic sink sink0 and writes it to the file f0.
>
> proc record {} {
>         global sink0  f0
>
>         set ns [Simulator instance]
>
>         #Set the time after which the procedure should be called again
>         set time 0.5
>
>         #How many bytes have been received by the traffic sink?
>         set bw0 [$sink0 set bytes_]
>
>         #Get the current time
>         set now [$ns now]
>
>         #Calculate the bandwidth (in MBit/s) and write it to the file
>         puts $f0 "$now [expr $bw0/$time*8/1000000]"
>
>         #Reset the bytes_ values on the traffic sink
>         $sink0 set bytes_ 0
>
>         #Re-schedule the procedure
>         $ns at [expr $now+$time] "record"
> }
>
> U NEED TO CALL THIS PROCEDURE INITIALLY IN THE MAIN PROGRAM.
> $ns at 0.0 "record"
> $ns at 100.0 "finish"  // to stop  the simulation.
>
> THEN THE FILE TO WHICH "f0" is the file descipter shows the o/p as
> 0.0 0.00
> 0.5 6.7
> 1.0 6.9
> 1.5 8.4
>
> --
>
> T.Chitti Babu (H-7, # 133)
> KReSIT , IITBombay
> MUMBAI-400076
>
> ---------------------------------------------
>
> INTERNAL PH NOS:
>         KReSIT Lab -- 5785
>         Hostel 7 No - 5607
>
> EXTERNAL PH NOS:
>         HOSTEL 7 Nos : 5721049, 5720105

--