[Mip4] Comments on draft-ietf-dhc-mipadvert-opt-01.txt

["Espen Klovning" <[e-mail]>]

Another late response to the poll on the list:

As Sami pointed out, the mechanism discussed in this draft has merit in more complicated foreign agent scenarios. I support such a mechanism.

Apart from all the other comments I have seen regarding this draft, I am still left with a few questions/comments. Any clarification will be appreciated.

1) I agree with Sami that a duplication of the existing agent advertisement format is not necessarily a good thing. Mostly because ICMP advertisements include more information than just the Mobility Agent Advertisment Extension, e.g. MAC addresses, ICMP Code fields, Prefix Length. The current extension is a good starting point though.

2) How is the MN supposed to use this mechanism? As far as I can tell, it is not mentioned at all. It might be intentional for all I know. So, should the MN:

A) Use the DHCP Mobility Agent Announcement directly without soliciting for an agent.
B) Solicit for agents but only accept the one(s) indicated in the Mobility Agent Announcement when the DHCP operation is completed.

There seems to be some issues with both approaches when the FA is not co-located but a separate physical entity.
Will the authentication implied by the draft still be useful? How will the MN get the necessary MAC address? Will the MN know whether the FA performs generic routing services or not? 

regards
espen 
Espen Klovning
Birdstep Technology ASA