Re: [radext] I18N help: : I-D Action: draft-ietf-radext-nai-02.txt

[Alan DeKok <[e-mail]>]

Pete Resnick wrote:
> This might be something we want to discuss in the contact of PRECIS, but
> let's get started just between us:

  I've been showing up at Precis for a while.

> Are these identifiers (usernames, realm names) things that users
> actually type in (i.e., user interface elements) or are they protocol
> elements? The answer will very much depend on which sort of thing they
> are. (And I have a sneaking suspicion that you're going to tell me that
> they are both, which is going to make your job much more difficult.) So,
> at what layer are these things used?

  A large percentage of the identifiers will be entered by users.  They
will then be used in various protocols to identify the users, and the
domains to which the users belong.

  The traditional practice in RADIUS has been to allow pretty much
anything as identifiers.  The only requirement is that the user and home
system agree on the user identifier.

  The problem we're trying to solve here is proxies.  They connect
visited networks to home networks.  The proxies need a standard format
for the user identifier.  This is so that they can route domain A to
network A, etc.  The format of the user portion is less relevant to
proxies.  It can be pretty much anything.  (and is, in practice)

  Alan DeKok.