Radio interface and Tree Discovery [wasRE: [nemo] NEMO multihoming]

["Pascal Thubert \(pthubert\)" <[e-mail]>]

Hi Vikrant:

As I understand it, 802.21 is defining a super radio layer (2.5 or less
:) that would perform intelligent handover over various radio
technologies.

At the moment, though:

L3 triggers (e.g. RAs in the NEMO case, or IGP messages) are ill/not
defined. There is no way for the access router to push them to the AP,
for APs to send the RAs as class1 on .11, or for the MN to know about
APs other then the one it is associated to. Even if that was available
(MN can get RAs from all APs in range), there's no control from L3 to
force an association to the best AP, in terms of reachability, from MN
standpoint.

Say there was, there's still a missing link for NEMO basic support to
operate efficiently. That's extended reachability information, inside
the RAs, besides plain ND. IMHO, there are 2 drafts that provide for
this in different contexts:

- For plain MN and multihoming, we have
draft-ietf-ipv6-router-selection-03.txt that gives us reachability
information in a routed topology.

- For Nested NEMO, we have draft-thubert-tree-discovery-01.txt that
enables the selection of the AR that builds the shallowest tree of
Mobile Routers to the Infrastructure (and much more :)

My thinking is that it's good for us to interact with 802.21 to get the
complete picture. But also and mostly, our responsibility is to move
forward with our own side of the 2.5 boundary, and in particular for
NEMO, there's work to do on the TD technology.

Pascal

> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]] On Behalf
Of Vikrant Kaulgud
> Sent: lundi 25 octobre 2004 06:21
> To: IETF NEMO WG
> Subject: [nemo] NEMO multihoming
> 
> Hi,
> 
> Ubiquity is one of the benefits mentioned in the draft and rightly so.
> Over the last few months, IEEE 802.21 has shown lot of activity in
> addressing media independent handoff between heterogeneous networks
> (http://www.ieee802.org/21)
> 
> Although 802.21 work is directed towards the MN, the basic concepts
> would still apply to multihoming in mobile networks also.
> 
> ~~~~~~~~
> Regards
> Vikrant