Re: Manual network access logins (Was: Re: [RAM] The mapping problem: rendezvous points?)

[Jari Arkko <[e-mail]>]

Brian, Tony,
>> I'm not an expert in this area, but isn't this type of thing already
>> addressed by 802.1x?

Yes, and 802.11i, EAP, RADIUS, built-in facilities in
802.16 and [23]G* link layers, many different optimizations
of the various parts of the process, and so on.

>>
>> Even if not, it's not hard to see the we will want to have automated
>> authentication mechanisms that allow us to change networks without
>> manual intervention.  
>
> But if the authentication is done for an ID instead of a locator,
> the AAA problem is somewhat transformed, I think.
>
>> So while we might not be there yet, this is a AAA problem that
>> someone should go off and solve.  The network architecture itself
>> should still allow for these transitions to be as seamless as
>> possible, including session migration.
>
> Indeed. Which, for real time apps, means that the latency for any
> AAA handoff needs to be pretty small. And maybe basing it on an
> invariant ID would help with that.

We need to be careful with the word identifier here. Currently,
network access is mostly based on either payment or subscriber
identity. But those identities are very different from IP layer
identifiers. Of course, the network access and IP layer setup
processes can be (and has been) tied together for optimization
reasons. Hopefully such optimizations do not create a 1-1
permanent binding between the subscriber identities and your
IP layer addressing or your future IP layer identifiers. This
would be bad for privacy reasons, for instance.

Jari


_______________________________________________
RAM mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/ram