[altq 328] Re: Efficient Networks cards

[Luigi Rizzo <luigi@info.iet.unipi.it>]

> Luigi Rizzo wrote:
> > but i don't understand, at least from what i read on this list you
> > can do shaping on ALTQ as well on the output interface ?
> > (and if the shaper is external, then dummynet does the job as well)
> 
> The difference is granularity.  Shapers on ATM cards provide cell
> level control driven by an oscillator so that it is quite accurate.

so, to understand again (as i said, i am pretty ignorant on ATM!) :
is the shaper part of the card's firmware ? I always thought it
was something built into the switches.

In any case, back to the problem: with altq you don't see the
cells so it does not really matter/make any measurable
difference to have cell vs. packet granularity, right ?
And for the kind of bandwidths mentioned (2Mbit/s) a cell will
go out every 200us, meaning that running with HZ=5000 might
give more than decent granularity for IP packets.

(I customarily run with HZ=1000 since i had 486/66, so i assume using 5000
on todays UHF processors should not be that terrible...

	cheers
	luigi

-----------------------------------+-------------------------------------
  Luigi RIZZO, luigi@iet.unipi.it  . Dip. di Ing. dell'Informazione
  http://www.iet.unipi.it/~luigi/  . Universita` di Pisa
  TEL/FAX: +39-050-568.533/522     . via Diotisalvi 2, 56126 PISA (Italy)
  Mobile   +39-347-0373137
-----------------------------------+-------------------------------------