Re: [IPsec] Call for independent experts (IKEv2) for Stage 4 of the PAKE selection process

[Tero Kivinen <[e-mail]>]

[removed cfrg from CC, as I do not think this issue really belongs
there as we are discussing IKE signaling here].

Dan Harkins writes:
>   First of all this suggestion is for a particular PAKE and I'm not
> suggesting that any of the other candidates would slide in so effortlessly.
> In fact an augmented PAKE is, IMHO, not suitable for a protocol like IKE
> where either side can initiate. The PAKE I'm describing here is SPEKE,
> a balance PAKE.
> 
>   SPEKE does a simple Diffie-Hellman but uses a secret generator that is
> deterministically obtained from the password. This technique is basically
> one of the hash-to-curve functions from the CFRG's hash-to-curve I-D
> or a simple hashing and exponentiation for MODP groups. All this happens
> at password provisioning time prior to IKE being run.
> 
>   Then when IKE is run the secret generator for the negotiated group is
> used to do the D-H, the IKE_SA_INIT exchange is basically SPEKE. The
> result is, if they both have the same generator (which means they had the
> same password), an authenticated shared secret. This secret is verified in
> the IKE_AUTH exchange.

How does the responder know which of the one million username password
pairs to pick to generate the generator when calculating D-H in the
IKE_SA_INIT? The actual identity of the user is only sent in the
encrypted IKE_AUTH message. 

I.e., I think this has exactly same problem than IKEv1 has with
pre-shared keys for main mode. You must know the initiator identity
based on the IP-addresses, thus makes this completely unusable for
non static VPN cases.

>   This would require a new Auth Method defined for SPEKE/PAKE to indicate
> that the SPEKE shared secret is used. And that should be all that's needed.
> It should be that simple. The protocol shouldn't have to change, no new
> messages, no new payloads, no new nuthin. If I'm missing something please
> let me know.
-- 
[e-mail]