Re: [Roll] do we need a dominating set?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Roll] do we need a dominating set?



Hello Maxence

 

During a sequence transition, the draft says that a child using the older sequence can use a ‘future’ parent with the newer sequence.

To avoid false positive in the loop detection mechanism, the rank is the Flow Label must be reset so the child appears as a host to its parent.

 

With what we discuss here,  say a  node that was router for sequence N-1 becomes non-router for sequence N. So the node never advertises N and all its children that see N should migrate over a transition period and stop using that parent.

Over the transition period, the node should still perform as a N-1 router, acting as prescribed in the draft today, that is resetting the rank in the flow label for all packets it issues or forwards.

 

Pascal

From: Maxence Dalmais [mailto:maxence.dalmais at insa-lyon.fr]
Sent: mardi 17 novembre 2009 09:26
To: Pascal Thubert (pthubert); roll ROLL
Subject: Re: [Roll] do we need a dominating set?

 

Hi Pascal,

the idea seems good to me but I got a question :
-How nodes act during the sequence spreading ?
I mean : Does a node which acted as a router and currently does not advertise as it continue to route during the T time ?
Moreover, does the fact to change his status of router or not does not increase packets lost or the number of sent ICMP Destination Unreachable packet ?

Maxence.


Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.