Re: MOBIKE WG Agenda for IETF-64, take 1

 In your previous mail you wrote:

   So, the authorization still happens the same way as
   defined in RFC 3776.

=> yes, the authorization is still needed.

   It is just that you are using
   a different mechanism to convey the home address i.e
   instead of ID/Traffic selectors i.e you are using
   MOBIKE mechanism.

=> I still use the traffic selectors but never the ID.

   Am i understanding this right ? 

=> more but perhaps not yet the whole idea. How to do express
the authorization in your favorite IKE implementation?
Does it work well with a dynamic HoA? With MOBIKE there is no
problem, or if your prefer, MOBIKE has the problems and
MIPv6 relies on the solutions used for MOBIKE.
   
   > So can we agree that to reopen the issue #7 is to
   > look at if/how to update transport mode SAs?

   At least that was my understanding :-)
   
   > BTW my opinion is the complexity is too high and the...

   I am not sure i agree on why this is more complex.
   
=> I have done many times the code for MIPv6 (updating tunnel endpoints
in IPsec SAs, SPDs, IKE) so I know it is not so difficult. I don't know
how to do for general transport mode, i.e, updating traffic selectors.
So I prefer to avoid a specification which is too hard to implement
(so won't be implemented) without a proof of a real need.
   
Regards

[email protected]