Re: Effective throughput

[Christoph Haenle <chris@cs.vu.nl>]

Ingela:

> trace file? I also thought maybe I can use a flow monitor somehow but I
> don't feel confident that parrivals and barrivals variables contain
> the effective throughput, maybe they contain the total throughput? I
> don't think the documentation is clear on this point so if someone
> would like to clarify it that would be great.  

parrivals_ and barrivals_ count the number of packets and bytes, respectively, 
that get routed to the link that is being monitored. To calculate the mean 
throughput of a link in bytes/s since the start of the simulation, you have to 
devide bdepartures_ by [$ns now]. bdepartures_ is generally smaller than 
barrivals_ because packets can cause the buffer of the router associated with 
the link to overflow (in which case barrivals_ is incremented, but not 
bdepartures_)

To reset parrivals, etc., just do a "$qmon reset".


     -Chris.



> 
> -- 
> /Ingela
> 
> //The highway of life is always under construction. //
> 
>  	        |\      _,,,--,,_  ,)
>                 /,`.-'`'   -,  ;-;;' 
>                |,4-  ) )-,_ ) /\
>               '---''(_/--' (_/-'