Responses in line...
On Nov 16, 2009, at 2:08 AM 11/16/09, Pascal Thubert (pthubert) wrote:
Hi Ralph
1. In a mesh-under network, the L2 characteristics of the lowpan are
close to those usually assumed for implementation of IPv6; in
particular, there are no "lowpan routers" at L3 and L3 messages
*can*
be delivered (perhaps with lower probability of success) directly
between lowpan nodes. Why is ND not sufficient in this model?
RFC 4861 ND works in a pure RFC 4944 mesh-under network if:
1) the mesh-under (L2) routing protocol provides subnet-wide
multicast,
2) that is efficient enough to be used for routine ND messages,
3) nodes are awake often enough to detect and reply to NS messages.
Such networks do exist, but these assumptions are not necessarily
compatible with many application
scenarios we have in mind; this is the reason we started with ND
optimizations.
Another way of saying this is that mesh under is more often than not
akin to NBMA. Fully mesh might exist and multicast emulation might
exist
though I've seen neither so far.
But there are actual networks out there where the model is really that
of virtual circuits, established either hub and spoke (to a sink) or
point to point (control loop).
So the model is closer to Frame Relay, for which as we all know 4861
is
not better suited than ARP was for IPv4. Issues:
* non transitive: A has a circuit to B and B has a circuit to C does
not
mean A has a circuit to C. Direct (one hop) multicast emulation cannot
do DAD.
* local addressing: A might not see B with the same address (VC nb)
as C
sees B. So A could not send redirects to C for B. And SLLAO is useless
(same as 3122) if VCs are switched.
OK, so these models really need to be documented so we can all start
on the same page. Which model(s) are you expecting for 6lowpan?
Our draft can act as a complement to 4861 in transit links (to grant
the
right to use an address) and a complement to 3122 on NBMA. Our draft
is
enough for host to router, and the existing ND operation still provide
value for host to host.
"host to router"? What address resolutions does ND-07 provide?
Would
proxy ND be sufficient?
I haven't seen such a design yet, so I don't know.
Recursing proxy ND demands a strict hierarchy / spanning tree. Could
be
done but probably difficult and inefficient.
This is certainly incapable to handle movement - there were drafts
like
this in NEMO long ago.
Why would recursing proxy ND be necessary? If the mesh-under provides
hub-and-spoke reachability, why not put proxy ND at the hub?
Cheers,
Pascal
- Ralph
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
