Re: [MEXT] Well-known problem with authentication/etc. in wireless networks

Hi Pete,

On Thu, Aug 25, 2011 at 11:40 AM, Pete McCann <[email protected]> wrote:
> Hi, Julien,
>
> Are you talking about EAP inside IKEv2?  That presupposes that the MN
> is already attached to the network somewhere and has an IP address (i.e.,
> it has already passed access authentication).

Yes, EAP authentication for IKEv2. Yes the MN needs to attach to the
network first, as hosts currently do today already.

> It may be interesting to look at whether access authentication and mobility
> management can be combined.

I don' t know what problem we would be solving by combining the two.

--julien

> On Thu, Aug 25, 2011 at 1:44 PM, Julien Laganier <[email protected]> wrote:
>> Charlie,
>>
>> I am not sure I understand what is missing in MIPv6; a MN and an HA
>> can already mutually authenticate using EAP, and this is incidentally
>> what 3GPP leverages on, together with the EAP-AKA method. What is
>> missing?
>>
>> --julien
>>
>> On Wed, Aug 24, 2011 at 12:06 PM, Charles E. Perkins
>> <[email protected]> wrote:
>>>
>>> Hello folks,
>>>
>>> It's now 2011.  Mobile IP was standardized late in
>>> 1996, after work had already been started nearly
>>> ten years before.  Over two decades! -- and regardless
>>> of lip service to fixed/mobile convergence we still
>>> don't have seamless mobility in user devices across
>>> heterogeneous media, and standards organizations
>>> (notably 3GPP) are not properly taking advantage of
>>> what Mobile IP can do. The losers are the end-users,
>>> which means all of us.
>>>
>>> There are many reasons for this, but one of the
>>> main reasons has to do with authentication at the
>>> access network.  EAP in various forms is being
>>> utilized for this purpose, and Mobile IP is not,
>>> even though there has never been any reported
>>> failure of the RFC 5944 or RFC 4285 or RFC 6275
>>> (to my knowledge).  Moreover, unless there is
>>> something wrong with the cryptography that also
>>> has not been reported, these authentication methods
>>> enable _mutual_ authentication between the network
>>> and the client, not just client authentication.
>>>
>>> In order for Mobile IP to enable the real promise
>>> of high performance heterogeneous networking, we
>>> have to do some more work.  I would like to initiate
>>> some more discussion about this.  DMM is interesting
>>> in its own right, but it's not at all the whole
>>> story.  Moreover, with proper design, it is likely
>>> the supposed burden of signaling to the home agent
>>> can be substantially reduced.  As one simple example,
>>> if handovers are accomplished locally between trusted
>>> access agents (routers, 802.11 access controllers, ...)
>>> then the actual timing of tunnel redirection from the
>>> home agent becomes much less critical.  This is also
>>> intricately intertwined with authentication.
>>>
>>> If the Home Agent were recognized as a robust security
>>> appliance, then it could naturally sit on the network
>>> boundary as an IP-addressable device.  Mobile IP
>>> authentication could become the primary means of
>>> validating user access, instead of an afterthought
>>> to enable IP-address preservation after all the heavy
>>> lifting has been done a lower levels.
>>>
>>> I would like to propose that in this working group we
>>> should go about making this happen.  It seems to be
>>> important, and undeniably aligned with our working
>>> group responsibilities.
>>>
>>> Regards,
>>> Charlie P.
>>>
>>>
>>> _______________________________________________
>>> MEXT mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/mext
>>>
>> _______________________________________________
>> MEXT mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/mext
>>
>