Hello ACE,

I have now addressed all WGLC comments (Jim Schaad's, Mike Jones' and Stefanie Gerdes') except for this one:

"Do we need to write something about how a RS should handle the presence of multiple tokens for the same client? Perhaps a security consideration?

I see two options:

1. Multiple tokens complement themselves i.e. if token A gives you right R1 and token B right R2 then you have R1+R2.

2. The newer token always overwrites the old one, which means if you want to extend your access rights as a client, when you already have A -> R1 you need to ask the AS for B*->R1+R2.
"
(see https://github.com/ace-wg/ace-oauth/issues/147).


AFAIK the common usage in OAuth is option 2, however Jim has pointed out use cases for option 1 and refers to it in
https://datatracker.ietf.org/doc/draft-schaad-cnf-cwt-id/


Jim has expressed a preference for 1. while Olaf has (in the Jabber at IETF 103) expressed a preference for 2.

I would need some guidance from the WG on how to proceed here.

/Ludwig

--
Ludwig Seitz, PhD
Security Lab, RISE
Phone +46(0)70-349 92 51

_______________________________________________
Ace mailing list
Ace@ietf.org
https://www.ietf.org/mailman/listinfo/ace

Reply via email to