Re: [Roll] Opinion please
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Roll] Opinion please
Hi all,
To eliminate children the parent can send DIO with infinite rank right
away? Or just resetting trickle, and starting T, which would expire
quickly.
Julien
> -----Original Message-----
> From: roll-bounces at ietf.org [mailto:roll-bounces at ietf.org] On
> Behalf Of Michael Richardson
> Sent: vendredi 6 novembre 2009 02:18
> To: roll WG
> Subject: Re: [Roll] Opinion please
>
>
> >>>>> "Jerald" == Jerald P Martocci
> <Jerald.P.Martocci at jci.com> writes:
> Jerald> I certainly favor the simplicity of Option 1 over
> Option 2;
> Jerald> however last week we also had an Option 3 which was my
> Jerald> favorite - that being that the node simply does
> not join the
> Jerald> DAG. Option 1 as described, indicates that the node joins
> Jerald> the DAG as a leaf node. My question is - how does a node
> Jerald> even know it's a leaf node? My understanding is that DAG
>
> mcr> It knows it's a left node because it knows that it failed
> mcr> to find a compatible OFy.
>
> Jerald> Why is this a discriminator? The node could have other
> Jerald> child nodes already attached to itself prior to the node
> Jerald> attaching onto the DAG. This set of nodes could
> all ascribe
> Jerald> to the same OFy, yet be different that the OFx of the main
> Jerald> DAG.
>
> okay, so if a node (AA) that implements OFy picks a new
> parent that has only OF0, it could then tell all it's
> children that it now has OF0, and those children, expecting
> OFy, would be in the same situation as AA.
> It can't forward OFx, because it really doesn't understand OFx.
>
> mcr> When it's a leaf node, it does not process the DIOs that
> mcr> it receives, and does not forward that information (list of
> mcr> prefixes, etc.) to it's parent.
>
> Jerald> Ok. I assumed that to eliminate newly found children the
> Jerald> node would stop sending DIOs. However, I wasn't sure if
> Jerald> DIOs are needed to be sent out periodically for other
> Jerald> reasons.
>
> I'm a bit vague on this myself. I think that it has to
> unicast DIOs to it's parent. It should not be multicasting
> DIOs if it doesn't want to be a parent.
>
> --
> ] He who is tired of Weird Al is tired of life!
> | firewalls [
> ] Michael Richardson, Sandelman Software Works, Ottawa, ON
> |net architect[
> ] mcr at sandelman.ottawa.on.ca
> http://www.sandelman.ottawa.on.ca/ |device driver[
> Kyoto Plus: watch the video
> <http://www.youtube.com/watch?v=kzx1ycLXQSE>
> then sign the petition.
> _______________________________________________
> Roll mailing list
> Roll at ietf.org
> https://www.ietf.org/mailman/listinfo/roll
>
Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.