Fw: [Click] Time and storage measurement

[Giorgio Calarco]

----- Original Message ----- 
From: "Giorgio Calarco" <gcalarco at deis.unibo.it>
To: "Paolo Pisati" <paolo.pisati at iit.cnr.it>
Sent: Wednesday, May 19, 2004 5:57 PM
Subject: Re: [Click] Time and storage measurement


> paolo
>
> if you are going to measure the end-to-end
> latency of your router seen as a black-box
> (i'm not sure if that is exclty your problem), you can:
>
> - use rude/crude (not so precise, not fast, but
> comfortable)
>
> - consider the maximum _loss-free_ forwarding
> rate. For instance, if you can forward 300Kpps,
> you need about 3.3 E-6 seconds for each
> packet (this comprises both the soft+ hardware
> delay)
>
> - use SetCycle + CycleCountAccum to measure
> the software delay and then make a theoretical
> evaluation of the hardware latencies required
> from your router (you need to be exactly
> conscious of what happens inside your PC and
> the kind of hardware you have)
>
> Obvioulsy, the latency varies if you change the
> packet size...
>
>
> I have no suggestions about the memory usage
>
> ciao
> giorgio
>
>
> ----- Original Message ----- 
> From: "Paolo Pisati" <paolo.pisati at iit.cnr.it>
> To: "Click" <click at amsterdam.lcs.mit.edu>
> Sent: Wednesday, May 19, 2004 4:19 PM
> Subject: [Click] Time and storage measurement
>
>
> >
> > Hi guys,
> >
> > this time i'm facing the task to measure the time my packets spend
> > while they pass through Click and, if possibile, it would be nice to
have
> > the memory usage of the entire router too.
> >
> > Something like:
> >
> > x ms spent in node A
> > y ms spent in node B
> > blablabla
> > ---------------------
> > Total: T ms spent in Click
> >
> > I don't think it's even possibile to do it in a NON-intrusive
> > way, so i'm all ears for any suggestions on how i should proceed to
> > achieve this.
> >
> > Thanks.
> >
> > Paolo
> >
> > -- 
> > Using Opera's revolutionary e-mail client: http://www.opera.com/m2/
> > _______________________________________________
> > click mailing list
> > click at amsterdam.lcs.mit.edu
> > https://amsterdam.lcs.mit.edu/mailman/listinfo/click