On 02/10/2017 19:46, Tero Kivinen wrote:
Jiliang Wang writes:
> Yes, there is channel scanning. Actually I have a question here. By
> current scanning mechanism, does it mean there is no guarantee for a
> scanning node and beaconing node to meet? Meanwhile, it may also take a
> long time to for two nodes to meet.Thanks for your explanation. I believe this is an important step in the entire protocol. The specification should avoid the scenario that the meeting time is too long or even unlimited (in any case) rather than depending on the implementation?
Yes. If you do scanning do fast or incorrectly it is possible that you
never hear EB. On the other hand if you do it so that you are sure to
hear the EBs it might take tens of minutes...
You can of course do all kind of implementation depending things to
make this better. For example if you do energy detection on the
channel first. If there is nobody ever on the channel for long time,
that might mean that channel is not part of the schedule, and
listening it might not be useful. On the other hand if there is too
much energy on the channel, it might be because there is some other
radio transmitter on the channel, and it might not be used for channel
hopping because of that reason. If you hear any 802.15.4 TSCH frames
on the channel, you can assume that this channel is in the current
channel hopping sequence, and if network is set properly there should
be EB on the channel at some point of time. On the other hand you
might need to wait for it for EB_PERIOD * slotframe size * timeslot
length * number of channels * n, where n depends how many dropped
beacons you want to wait. I.e. with typical settings this might be few
minutes. If you do not see EB during that time, you move to next
channel and repeat...
--
[email protected]
_______________________________________________ 6tisch mailing list [email protected] https://www.ietf.org/mailman/listinfo/6tisch
