On Wed, 22 May 2002, Jari Arkko wrote:
> I do agree about the correspondent node functionality part, though. However,
> there's a couple of things we should observe. In the current design,
> correspondent nodes do not need do anything special unless they want to
> do Route Optimization. The exact keyword for Route Optimization is also being
> debated. My interpretation of the discussion is that SHOULD is winning.
> In any case, communications with e.g. old IPv6 nodes that do not yet
> support MIPv6 are always possible without any other problems than
> non-optimal routing.

Are you sure these are the only problems?  It seems to me, that these old
IPv6 nodes's, ie. no support for even HAO, cannot even talk to mobile
nodes until receiving an ICMP message (one roundtrip).

That's because Destination Option code for HAO is 209 -- if the 
destination node does not recognize it, the packet is discarded and an 
ICMP message sent.

Actually, I'm not 100% sure from MIPv6 -17 draft that MN's are to process
and take account ICMP parameter problem code 2 messages sent by these old
nodes, in any way.

(Even though this wouldn't be a problem, additional cost of zero support 
in CN's would be that connections would break if MN moves.)

Perhaps I missed something?

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords

--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to