Re: [nemo] RE: draft-ietf-nemo-home-network-models-05

[Alexandru Petrescu <[e-mail]>]

Ryuji Wakikawa wrote:
 > Alex
 >
 > On 2006/02/17, at 1:37, Alexandru Petrescu wrote:
 >
 >> Ryuji Wakikawa wrote:
 >>> Hi Alex and Thierry Please check the RFC3964 section3 We said "A
 >>>  Mobile Router has a unique Home Address through which it is 
reachable when it is registered with its Home Agent.  The Home Address 
is configured from a prefix aggregated and advertised by
 >>>  its Home Agent.  The prefix could be either the prefix
 >>> advertised on the home link or the prefix delegated to the Mobile
 >>> Router. "
 >>>
 >>
 >> Hi Ryuji, there can be no implementation that derives the HoA from
 >>  MNP and simultaneously is defended by proxy ND by HA.  Is there any?
 >
 > I don't know how you can conclude like this..
 >
 > SHISA can generate HoA from MNP. MR can return home in some 
configuration. When MR returns home, HA does not remove network route
 >  of MNP pointed to MR. Thus, without proxy ND, MR can return home.

Mobile IPv6 and NEMO require HA to do proxy ND for the Home Address.  If
some HA does not do proxy ND for the Home Address then it's a little bit
incomplete.

That's why I was saying that one can't implement a Mobile IPv6/NEMO HA
doing proxy ND for a Home Address derived from the MNP.

 >>> Some advantages using HoA derived from MNP. The operators do not 
suffer from matching HoA from the home link and assigned MNP. We can 
save a prefix/64 for just assigning HoA to MRs.
 >>
 >> I'm not sure I understand how the matching hurts the operators.
 >>
 >> I don't think addresses or prefixes are saved, can you explain?
 >
 > You need /64 or shorter prefix to assign HoA.

You need a /64 prefix for all the MRs and MHs (and maybe some LFNs too)
present on the home link.  It is at least needed to put something in the
RAs the HA sends on its home link (otherwise MR can't realize it is back
home).

draft-xx:
 > To properly identify the Home Network, the MR must expect a shorter 
prefix than that of the Mobile Network from which the Home Address was 
formed.

This is not implementable.

Most prefix match in routing is a "longest-match".  A shorter prefix
can't be really matched  unless a precise bound is specified, but then
it's an "exact match".

I.e. if a MR in "aggregated mode" attaches to a network whose first two
bits match its Home Address, but all the other (3rd up to 63rd) bits
don't match, will falsely believe it is at home.

Alex