eap Mailing List: Re: Clarifications on "Domino effect"

[Jari Arkko (jari.arkko]

Going through the still open issues...

> In RFC4017, the Domino effect is described:
>  
This text is also in keying-08.

> "Compromise of a single authenticator cannot compromise any other part
> of the system, including session keys and long-term secrets."
>
> Does this finally imply that the authenticator MUST not provide keys to
> other entity ?
>  
Intuitively, this text may be too broad. An EAP peer, for instance,
is a part of "the system", and its traffic at least would be
compromised if its authenticator got compromised. And any
new EAP peer connecting to the compromised authenticator
would also have its traffic exposed. Similarly, the AAA nodes
are affected, because they have a secure connection to a
compromised node.

I think what we mean is that when one authenticator is
compromised, this does not lead to:

o  Compromise of long-term secrets in EAP peers, AAA servers,
   and other authenticators.

o  Compromise of session keys other than those associated
   with a session where the compromised authenticator
   is or will be a part of.

o  Ability of the authenticator to claim to be another authenticator
   or to offer another type of service when communicating with
   EAP peers.

--Jari