On Nov 1, 2010, at 2:03 PM, Omprakash Gnawali wrote:

> On Mon, Nov 1, 2010 at 2:52 AM, wasif masood <[email protected]> wrote:
>> 
>> Yes, I think it would be troublesome if one neighbor completely vanishes
>> from the vicinity of a node because in that case node will always be
>> transmitting 3 byte additional overhead in terms of the reverse link
>> neighbour quality.
> 
> Good point! If you want to fix that, you should expire the
> entry after a certain number of maximum Trickle intervals. I don't
> know how big of a problem this is so I am not convinced yet if we
> should make this change.

In a network with sufficient density, the dead neighbor should be replace by a 
not-dead one. So the question is whether the (tiny) energy savings are worth 
the increase in code size to handle this edge case. My guess is no.

Phil
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to