[Tinyos-devel] 2.0.2 CC2420

[Hugo Sousa]

I don't mean to intrude as I haven't been following closely this
discussion, but I must also say that I agree with J. Ryan. In my case
I use it for a particular feature in a time synchronization protocol,
when a mote powers on it asks the network (broadcast) what time it is
and whoever receives it answers directly (unicast) to the mote. To
prevent a flood of answers they (randomly) delay their response and
snoop for others answering, and cancel their response if someone else
already answered.

I don't know if this is a stupid suggestion, but what if the
instantiation of an AMSnooperC/SnoopingReceiverC would disable the hw
check? (the #define would be inside the AMSnooperC/SnoopingReceiverC,
and the user wouldn't need to know what to do - maybe just a
#warning). If the users would still prefer the SACKs and software
check they could #define themselves that

Best regards,
Hugo Sousa



On 7/4/07, J. Ryan Stinnett <jryans at rice.edu> wrote:
> David Moss wrote:
> > There are two settings for acknowledgements in CC2420ControlP:
> > enable/disable auto acknowledgements, and make hardware or software
> > acknowledgements the default.  These can be configured, as they were before,
> > with preprocessor #defines, and now can also be configured through a
> > software interface (CC2420Config) at runtime.  CC2420_NO_ACKNOWLEDGEMENTS
> > disables ack's completely, which is what the base station has set in its
> > Makefile.  So, the base station will continue to sniff packets because its
> > address recognition and auto acknowledgements are disabled.
> >
> > Since the software address check would have only affected the SACK being
> > issued, it should have no effect on the base station.  If you want the base
> > station to disable address recognition to sniff all packets, and also send
> > out SACK's when it receives a packet with its address, that's another issue.
>
> This sounds like a useful feature to me, not for the base station
> necessarily, but for general motes.  I'm currently working on a routing
> protocol which can make use of data from overheard packets, but I don't
> want to give up the acknowledgment functionality in order to also snoop
> packets.  For my current application, this would be the optimal setup,
> so I wouldn't call it an edge case at all.
>