ORBIT-USER: Questions about MAC cards ...

[Haris Kremo]

My concern is whether TxOp is turned on by default.

If you see uninterrupted bursts of the same number of packets this might
be the reason.

The basic protocol is designed in a way to, at least try to, provide
equal opportunity to all nodes to transmit. In practice this may not be
the case, for instance due to signal strength differences, and one can
expect unequal contribution of the nodes to the aggregate throughput -
but in a more random way.

Hope it helps,

H.

On 12/3/06, Brian K. Russell <morbius at cs.rutgers.edu> wrote:
> On Sat, 02 Dec 2006 18:04:14 -0500
>  Mesut Ali Ergin <ergin at eden.rutgers.edu> wrote:
> > Assuming you are referring to wired/wireless NICs,
>
> Yes, that is true.
>
> >
> > 1. They do have multiple frame holding buffers. (e.g. ORBIT Atheros
> > 5212 cards have a 50-frame buffer)
> >
> > 2. They behave exactly the way an interrupt driven device should 
> > behave. You can give it as many as 50 packets as fast as you can, it

> > will let you know what happened to those packets (individually or in
> > bunches) by generating interrupts and calling registered ISR of the 
> > driver.
>
> This explains a great deal.  If I generate packets at some fixed rate 
> and send them to the buffered MAC card, they may not get transmitted 
> at the same rate.  In fact, under conditions of contention, I may end 
> up buffering all 50 packets and then having them go out in rapid 
> succession.  That would explain the burstiness I've seen quite nicely.
>
> Thank you.
>
> BKR
> >
> > --
> > MAE
> >
> >
> > Brian K. Russell wrote:
> > > 1. Do MAC cards have buffers for multiple outgoing packets or is
> > there
> > > room for just one packet at a time?
> > >
> > > 2. If a MAC card is busy with an outgoing packet (unicast or
> > broadcast),
> > > what happens if another outgoing packet is given to the MAC card
> > before
> > > it's finished with the its current outgoing packet?
> > >
> > > Thank you.
> > >
> > > BKR
> > >
>
>
>