>>>>> On Thu, 21 Mar 2002 08:17:06 +0900 (JST),
>>>>> Keiichi SHIMA <[EMAIL PROTECTED]> said:
>> 1. revise the mobility section to clarify issues about using
>> site-local addresses with mobile IPv6:
>>
>> => I disagree with your solution because it takes more than a few line
>> to describe it (:-).
>> IMHO site-local addresses in a mobile IPv6 work well if only a
>> bidirectional tunnel is used between the mobile and its home agent.
>> Note there should be no penalty because by definition communications
>> in the same site are local, so the only difference with routing
>> optimization is the encapsulaion overhead.
>> Another important point is an equivalent way to describe this solution
>> is to say the bidirectional tunnel belongs to the home site.
> Are you saying that if we use a bi-directional tunneling, the off-site
> MN can comunicate with the CN that is in the home-site of the MN?
> If so, I disagree. Consider the following condition.
> site A (fec0:0:0:100::/64): a home site of the MN
> site B (fec0:0:0:100::/64): a foreign site.
> If the MN moves from the site A to site B,
In this case, the mobile node is multi-sited. The following figure
would depict the situation.
+----------------------------------------+
| +=====tunnel link====================+ |
+ | +--------------------------------+| |
\| / || |
I1 +--------+| +--home site-----+
---- MN ---foreign site--+ | HA |
I2 | | |
The MN has (conceptually) at least two interfaces; I1 and I2. The
former (conceptually) belongs to the home site, and the latter of
course belongs to the foreign site. The site-local home address
belongs to I1 (or perhaps a different interface, but it should belong
to the home site).
Just like other (normal) multi-sited nodes, the MN should always
disambiguate the site zone when sending a packet. This applies
when the mobile node sends a packet to a node in the home site.
So,
> the MN cannot determine the
> direction to which it should send a packet destinated to
> fec0:0:0:100:1234:5678:9abc:def0. The destination address can be both
> on the site A and the site B.
this is true, but cannot be a counter argument to what Francis said.
We should (and can) just disambiguate the site zone when sending.
BTW: It's still true that this type of situation is problematic.
For example,
- there is no way for the mobile to detect if the site A and the site
B are same or different site (as I said in my first message in this
thread).
- if the site A and the site B are different sites, the mobile node
cannot communicate with a node in the site B using mobile IPv6.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
