Cell Relay Archive

[albert.e.manfredi@boeing.com]

In article <34FD5172.AF30FE0A@intec.rug.ac.be>,
  Mike Vogeleer <Mike.Vogeleer@intec.rug.ac.be> wrote:
>
> Hal Murray wrote:
> >
> > In article <34FC23D0.542BF517@intec.rug.ac.be>, Mike Vogeleer
<Mike.Vogeleer@intec.rug.ac.be> writes:
> >
> > > Does anybody know the typical value for the queuing delay of an
unloaded
> > > ATM switch. With 'queueing delay' I mean the time that elapses between
> > > having received the last byte of a given cell on the incoming interface
> > > and starting to send the first byte of that same cell on the
appropriate
> > > outgoing interface.
> >
> > For the GIGASwitch/ATM it's roughly 15 microseconds per hop.
> > [That's the first-byte-in to first-byte-out time.]
>
> I really meant last-byte-in to first-byte-out time. At 622Mbps and
> 53-byte packets this should be 15 microsec - 53*8/622 microsec = 12
> microsec from the last-byte-in to the first-byte-out. Correct?

Are you assuming that the entire cell is buffered into the switch _before_
the first byte is allowed out the other side? Maybe that's the problem. Only
the cell header needs to be checked.

Bert
manfredi@arl.bna.boeing.com

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/   Now offering spam-free web-based newsreading