[VulnDiscuss] Re: Cracking preshared keys

Damir Rajnovic wrote:

>Please note that the same class of attack is possible even if 
>Xauth (Extended Authentication) is used. This is because Xauth is 
>performed after Phase 1 is completed and, for this attack, an adversary
>needs only a packet from Phase 1. Furthermore, after the pre-shared 
>key has been discovered, an adversary can mount an active MiTM attack
>on Xauth. The outcome depends on the exact authentication method used
>in Xauth.
>
May I ask how this applies to "IKE Shared Secret AAA"?
If a Radius backend authenticator is used, is the shared key
that is vulnerable one preconfigured in the VPN hosts or one
based on the Radius password? The reason I ask is that the
preconfigured shared key can easily be lengthened but the
backend password, if based on end user passwords, is not
as easy a solution. :)

Thank you.

Gary Flynn
Security Engineer - James Madison University