Re: [radext] Adoption call for draft-perez-radext-radius-fragmentation-06

[Alan DeKok <[e-mail]>]

Stefan Winter wrote:
> That's not the end of the world though; simply adding a sentence in the
> draft such a case exists and that deployments need to take care that
> their deployed implementation(s) don't stumble over this.

  Yes.  Even if there are deployed proxies which implement 6929, there
are no attributes which use the 6929 format.  So there's a bit of time
to upgrade the systems.

> Aha! I had the impression that authz exchanges in the ABFAB world go all
> the way from/to the RADIUS client that does the whole GSS EAP magic. And
> that this is the equivalent of an "AP" in the non-network-access use case.

  In the abfab world, they may go to the client, which can be a
commodity system with user logins, web browser, etc.  i.e. a capable system.

  I don't recall seeing a demand for APs or switches to implement the
fragmentation draft.

> That's something I can live with. Again, documenting in the draft that
> User-Name is the only proxying criterion would be enough for me then.
> Maybe a second one explaining that client to first-server is not in
> scope/has complexities that can make the use of this draft hard/unsuitable.

  Yes.  We'll do that.

  Alan DeKok.