Re: [radext] [precis] saslprepbis and NAI

[Peter Saint-Andre - &yet <[e-mail]>]

On 11/14/14, 7:17 AM, Alan DeKok wrote:
>    The RADEXT WG is in the last stages of publishing the updated NAI
> document:
>
> https://tools.ietf.org/html/draft-ietf-radext-nai-10
>
>    This defines a standard for "username@realm" handling, where "realm"
> is a fully qualified hostname.  It would be useful to reference that
> from the saslprepbis document.

What kind of reference do you have in mind? Perhaps a pointer from 
Section 3.4 ("Application-Layer Constructs")?

>    The NAI is used for EAP (802.1X), AAA, 3GPP, and international roaming.
>
>    The NAI is a subset of the "userpart" defined in Section 3.1 of the
> saslprepbis document, in that it *requires* "username@realm", and the
> "realm" part is a fully qualified domain name.
>
>    Hmm... I've just noticed that the ABNF in the NAI allows for IP
> addresses in the "realm" portion.  That's not the intent.  I'll have to
> go fix that.
>
>    I'd like to be sure that the NAI document is compatible with (or at
> least not in violent disagreement with) the precis work.

I've just reviewed the NAI specification. Although it does directly not 
address issues such as width mapping, case mapping, and Unicode 
normalization (seemingly inheriting its handling of those issues from 
RFC 5335 for the username portion and from IDNA2008 for the realm 
portion), as far as I can see it is compatible with saslprepbis.

Peter

-- 
Peter Saint-Andre
https://andyet.com/