Pascal,
Here are some comments on backbone-router-01:
First, this nicely integrates Erik's 6lowpan-reg-00 concept in a way.
The Binding Solicitation message is equivalent to the ND Registration
message. Instead of piggybacking on the RA, BbR uses the explicit
Binding Confirmation message. However in BbR the process is opposite to
that in 6lowpan-reg. I prefer the model used in the BbR draft, where
binding maintenance is controlled by hosts. The forced registration to
multiple routers at a constant interval seems to be an overkill to me in
6lowpan-reg.
I like the optional secondary binding concept, which is controlled by
the host. This is a good approach.
In Section 5.1, paragraph 3, "optimistic node" should be "optimistic mode".
Section 6.1 has a question about the use of RAs, and finding the BbR
faster when arriving to a new LoWPAN. I would actually say that
listening for RAs, sending an RS, or sending a BS to 6LOWPAN_BBR are the
only L2 independent ways. In case you have a LoWPAN running in ad-hoc
mode then this is the only way. Of course if joining a beacon-mode
network tells you who the PAN-coordinator (and thus the BbR) is quickly
at L2, then it doesn't hurt to use that info.
If I understand right, BbR also allows for the use of stateful
compression from bormann-6lowpan-cbhc-00?
Reading the draft carefully, I don't see anything preventing the use of
mesh under or route over. The only issue is that you must (should?)
route all traffic through your primary BbR, thus when using route over
you should indicate the primary BbR as an intermediate hop. Or can you
send traffic outside the lowpan through a secondary BbR at will? In 6.5
it states that as long as a binding entry exists for the source address,
then forwarding can occur. This requirement should be explained in the
draft better.
In my opinion the BbR proposal solves lots of open problems and
questions in chakrabarti-6lowpan-ipv6-nd-05, while almost completely
avoiding the use of multicast and efficiently using the IPv6 router to
offload ND operations. Very useful in practical networks. The timeline
for ND optimization submission is in Nov, it would be ideal to
eventually merge the BbR draft into the ND optimization draft?
- Zach
--
Zach Shelby | Head of Research | +358 40 7796297
Sensinode Ltd. www.sensinode.com
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
