Re: [MEXT] Reviews of draft-korhonen-mext-mip6-altsec

[marcelo bagnulo braun <[e-mail]>]

  I agree with all Julien stated.
Just one more comment.

I personally would be interested to understand for each new proposed 
solution, what would be the interoperability plan with the current 
solutions. I mean, it could be as simple as: " we assume that all nodes 
in the network implement the new solution", but maybe there is something 
more into it (e.g. the HA will support multiple security mechanisms, and 
the MNs will support only one).
I think it would be worth doing the exercise.

Regards, marcelo

El 14/10/10 19:04, Laganier, Julien escribió:
> Sri,
>
> Sri Gundavelli wrote:
>> Marcelo/Julien:
>>
>> Clarifying question ?
> Sure.
>
>> Given that, this (Alternative Security Mechanisms for MIPv6) is now a
>> chartered work item under experimental category, wondering about the
>> future and the relation of the candidate solution with two existing security
>> mechanisms, before we spin one more.
> To be exact: we do not spin one more, we are spinning more experimental mechanisms, to experiment with security model alternatives. There might be one more, two more, etc. It all depends on what the WG think is worth experimenting with.
>
>> We have IPsec, a normative standard, Authentication Option, an
>> informative standard, and hopefully a perfect new solution as an experimental
>> standard.
> To be exact: RFC 4285 (Authentication option) does not specify an Internet standard of any kind.
>
>> So, what is the evolution plan, if we love this new experimental
>> standard and MIP deployments get adopted and phase out GTP in 3GPP, will this
>> get promoted to be an informational standard or a normative standard ? :)
> Any new experimental RFC that the WG publish will define an Experimental Protocol for the Internet community.  It will not specify an Internet standard of any kind.
>
> As a result of the experiment, if one of the proposals is considered successful we might consider progressing it on the standard track.
>
>> Secondly, for Auth-Opt to be considered an alternative security
>> mechanism, since that is some what black listed with that famous IESG note, will
>> it be a demotion or a promotion to get that under this Experimental standard,
>> from informational standard with a red dot. :)
> Since neither experimental RFCs not informational RFC defines Internet standard or any kind, there does not seem to be a huge difference. However from my perspective the experimental status highlights that an experiment is ongoing and thus that at some point the results can be evaluated by the community.
>
>> Just curious, we will surely have excellent interoperability between
>> vendors. Its better to put few forward looking statements around this
>> work, else it will confuse the hell out of every one.
> Standard wise there is no reason for anyone to be confused. IPsec remains the standard track, mandatory to implement security mechanism that guarantees interoperability.
>
> --julien
>