On Nov 2, 2011, at 9:20 AM, Alejandro Perez Mendez wrote:

> 
>>>> "Rhys" == Rhys Smith<[email protected]>  writes:
>> 
>>     Rhys>  The use of ABFAB technologies in this case, via a mechanism
>>     Rhys>  dubbed "federated cross-layer access" (see [ref]) would
>>     Rhys>  enhance the user experience of using these applications on
>>     Rhys>  mobile devices greatly. Federated cross-layer access would
>>     Rhys>  make use of the initial mutual authentication between device
>>     Rhys>  and network to enable subsequent authentication and
>>     Rhys>  authorisation to happen in a seamless manner for the user of
>>     Rhys>  that device authenticating to applications.
>> 
>> 
>> I'm not fond of the architecture of re-using the authentication or at
>> least am nervous about the architectural coupling involved.  Why isn't
>> it sufficient to simply re-use the credential? I.E. we cann't we perform
>> a second mutual authentication.?
> 
> Well, I guess that the point is exactly that, to avoid that second mutual 
> authentication by resuing cryptographic material generated during the network 
> authentication. An average user would like to enter his PIN number only once, 
> when the device starts, and then, he can access services seamlessly, without 
> being prompted about the PIN number again. This is even more interesting if 
> instead a PIN number the user has to write a passphrase (not thinking in a 
> mobile phone now, but in a WPA-enterprise authentication, for example).
> 
> If you are thinking on providing SSO within all the services of the 
> federation, as we do in our kerberized proposal, the intention is to 
> incorporate the network access as another service, without worrying in which 
> layer it works. As it is usually the first service accesed by the user in a 
> session, it would be in charge to "bootstrapp" the SSO.

yes, I agree with that. Furthermore, I can imagine mobile operators offering 
brokered services (operator A strikes a deal with SaaS provider B) to their 
users, by definition the authentication process of the operator is the one you 
want to rely on. I could also imagine a step-up authentication sort of 
scenario, in which the network authentication acts as the first factor and 
another authentication method is used as additional authentication for services 
that have a demand for stronger authentication.

Klaas

P.S. I like Rhys' wording

> 
> Regards,
> Alejandro
> 
>> 
>> --Sam
>> _______________________________________________
>> abfab mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/abfab
> _______________________________________________
> abfab mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/abfab

_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab

Reply via email to