Hi Carsten:

Fact is, the latest 6LoWPAN ND draft does not support mesh under since
it lost the white board piece.
It does not support route over either since it is not compatible with
the only route over protocol in existence.

That's very far from what the group voted in Dublin when it elected the
backbone router draft as WG doc.

At the moment, I see 2 possible directions for 6LoWPAN ND:
- Pack it an Indiana Jones RFC and hide it in the midst of 6000 other
boxes where it will be mostly harmless.
- Roll back to where we were before Hiroshima and work closely with RPL
to build a consistent solution for route over.

The ball is in your camp.

Pascal


> -----Original Message-----
> From: Carsten Bormann [mailto:[email protected]]
> Sent: Wednesday, May 05, 2010 5:44 PM
> To: Mathilde Durvy (mdurvy)
> Cc: Pascal Thubert (pthubert); [email protected]; Richard
Kelsey;
> [email protected]; [email protected]
> Subject: Re: [6lowpan] [Roll] how does a node get an IP address
> 
> > On a related topic....
> > 6lowpan network have the particularity that you cannot use on-link
> > prefix due to the non-transitivity of the wireless links. This means
> > we need to tell routers how to reach neighboring IPv6 hosts. So
> > essentially 6lowpan-ND is using a registration mechanism to
establish
> > a "route" between the router and the host.
> > It is not clear to me whether this is the role of ND
> 
> The role of ND here is doing the host-router neighbor discovery.
> 
> > or of the routing
> > protocol.
> 
> The role of RPL here is to disseminate the host route on towards other
> routers.
> 
> > I think it could actually be both.
> 
> Hosts should not need to speak routing protocols.
> 
> > Hence the questions:
> > - Are IPv6 hosts possible in a 6lowpan network where the RPL
protocol
> > is used?
> 
> I would consider it a major failure if RPL didn't support networks
with hosts.
> (Networks without hosts, i.e., all-router networks, are certainly
possible, but
> in a constrained node/network we are talking about a relatively high
lower
> bound on the complexity of nodes or a *very* simple routing protocol.
> Of course, the routing protocol could distinguish three instead of two
> complexity classes of nodes, supporting something like a "stub router"
[a
> router that cannot forward].  Still, that class would need to
implement some
> parts of the routing protocol, imposing constraints either way or
both.)
> 
> > - Should IPv6 hosts be part of a RPL topology (as leaf node) or
should
> > IPv6 hosts use the 6lowpan-ND host-router spec?
> 
> As I said, hosts shouldn't need to speak RPL.
> 
> Gruesse, Carsten

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

Reply via email to