Sorry Jerry I was not clear ... I just wanted to clarify that the term multicast is used in an IP context (not referring to the link layer).
Hi JP,
I don't think RPL can simply state 'IP Multicast' as you note below. As Michael mentions there are a) unicasts, b) link-level multicast and c) subnet-level multicasts. He further mentions that on Ethernet link-level and subnet level multicasts are the same. I totally agree with this. However, the converse of this is not true; on wireless mesh networks link-level multicast and subnet-level multicasts are very different.
On a 6LowPAN based mesh network for example the link-level multcast domain will be simply the nodes within direct radio range of the source; while a subnet level multicast domain will cover the entire LLN. I believe RPL has to clearly state the extent of the multicast domain in every case; else different implementations may define different multicast domains yielding no interoperability. Furthermore the network traffic patterns will be hugely impacted if multicasts cross the entire LLN, when the algorithm was meant to be limited to the source's neighbors.
Jerry
On Nov 16, 2009, at 8:50 PM, Michael Richardson wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
>>>>>> "Jerald" == Jerald P Martocci <Jerald.P.Martocci at jci.com> writes:
> Jerald> I agree with Julien that we need a DIO solicitation message
> Jerald> (i.e. DIS) since as he says a packet can't wait around for
> Jerald> hours waiting for a new DAG parent. I buy that and agree
> Jerald> with it. However, I am a bit confused why a multicast DIS
> Jerald> should trigger a multicast DIO. This seems to simply add
> Jerald> unnecessary packets onto the network.
>
> I am reacting to the last sentence.
> I'd have thought that multicast avoids unncesary packets!!!!
> So perhaps we are not talking about the same thing?
>
> Maybe we need better terminology.
> I think that there are three kinds of transmission:
> a) unicast
> b) link-level multicast
> c) subnet-level multicast
> On ethernet, (b) and (c) are the same thing.
>
Just bear in mind that we always refer to IP multicast, since we are
not link layer specific.
> On 6lowpan-type networks, I think (b) is what happens when you
> transmit to the nodes that can hear you, and (c) is what you do via a
> whiteboard or multicast relay.
>
> (Maybe there are accepted terms for this, of which I am ignorant)
>
> Only transmissions of type (c) cause more packets on the wire.
>
> Transmissions of type (b) would save packets, because one message
> arrives at many receivers at the same time. Maybe (b) causes
> receivers
> that would otherwise have been happy to sleep (doze?) to wake up.
>
> - --
> ] 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.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (GNU/Linux)
> Comment: Finger me for keys
>
> iQEVAwUBSwGs6YCLcPvd0N1lAQKt/QgAoaq8+YxFPjcDwbZMyl5lfHHpvr+khEdH
> LgGd+8AmdNNfeiINH/369a02FgMuAZQjb1SHGdGsZZcR0JcB5BajalaAJy3VD0yL
> IwfUxGwlaHZrcmJaBkQ5ShHOH4tUizo/9hLar4vEtFu4EGLUaRR5C86VMTdtLs1s
> eUXiDLXhaftjyVhDC8TsYOY92FGdw/cmy8LFXNgCSBVqPgUwOV5t4oo5lngvzhQp
> U5Znba6IvrHmMb4w8m8UIkuKZ1ugaBBQG9/S2a5EAAw+6vBK1x6aHVnxpdQXS5qG
> VnCoxQlYIS5aBY/zPl+/siB8srLkEpKUHlTquVH8Ee0QgncNgdftqg==
> =MCIA
> -----END PGP SIGNATURE-----
> _______________________________________________
> Roll mailing list
> Roll at ietf.org
> https://www.ietf.org/mailman/listinfo/roll
_______________________________________________
Roll mailing list
Roll at ietf.org
https://www.ietf.org/mailman/listinfo/roll
_______________________________________________
Roll mailing list
Roll at ietf.org
https://www.ietf.org/mailman/listinfo/roll
|
