On Mar 20 2008, at 13:21, Porcu Giorgio wrote:

> But what happens within a PAN when a backbone router crashes? How  
> can a =
> node get routing info about other nodes belonging to its same PAN if  
> it =
> can't access the white board residing on the crashed backbone  
> router? =

I'd say: The "backbone router" as a SPOF is a problem *if* the network  
can do anything useful without the "backbone router".

If it can't (e.g., the whole point is connecting the sensors to some  
other node on the Internet), then the network can just fall apart (of  
course, the failure still needs to be battery-efficient, and there  
needs to be an orderly re-bootstrap when the/a backbone router comes  
up again).

If it can: you need to elect a new coordinator.  That might work best  
if there is a backup coordinator in operation all the time.
Multiple coordinator candidates do cause all the partition/merge  
problems that I won't go into right now.

A coordinator is also a useful thing to have for key management, so I  
believe it is best to think of the address management and key  
management problems in combination.

Gruesse, Carsten

PS.: Giorgio: You need to fix your mail setup: you should never send  
ms-tnef towards the Internet; I'm not sure how many people could even  
read your message.

_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to