Re: [netext] Review of I-D: draft-ietf-netext-pmip-lr-04

[Qin Wu <[e-mail]>]

----- Original Message ----- 
From: <[e-mail]>
To: <[e-mail]>
Cc: <[e-mail]>; <[e-mail]>
Sent: Friday, September 16, 2011 6:11 AM
Subject: Re: [netext] Review of I-D: draft-ietf-netext-pmip-lr-04


 >>> It would be useful to at least provide some examples of the tunnel
>>> 
>>> establishment between MAGs from a completeness perspective. It looks
>>> 
>>> like handwaving at the moment.
>>
>>I agree with you that it is handwavy, but I am not sure what to add
>>here. People wanted all kinds of tunnels here, IPv6, IPv4, GRE, USP,
>>IPsec etc. with either manual or dynamic creation. If you want a
>>specific scenario, I could add it here but if I remember correctly we
>>did not even manage to get consensus among the authors.
> 
> Without actually specifying at least one default tunneling mechanism, how
> would you achieve interoperability between MAGs in a PMIP6 domain? Whether
> it is IP-in-IP or GRE, it does not matter. I would recommend that the I-D
> specify the details about the tunneling between the MAGs. I hope you can
> discuss and get consensus among the authors and from WG members.

[Qin]: I think the default tunneling mechanism should be IP-in-IP, just like what it said
in the section 6.10.2 of RFC5213. e.g., RFC2473 can be used to establish IPv6-in-IPv6 tunneling.
If additional tunneling mechanisms is introduced in, Negotiation
on which tunnel mechanism is selected is required, however as described in RFC6279, this should be 
out scope of netext. 
If dynamic mechanism is used to establish tunneling, we should follow RFC6279 to establish localized
routing states on relevant MAGs.

The question is is it enough to use LRI/LRA to populate localized routing 
states to the MAGs? Do we need addtional extension to LRI/LRA
used between MAGs?