> > How does EAP/MD5 compare with EAP/TLS in regards to data 
> encryption over the air between the 
> > Client and the AP?  Is it the
> > same, just using a different mechanism to do the encryption 
> (MD5 challenge, verses a 
> > Certificate)?
> 
> Short answer: 
> 
> NO, it's not the same! EAP/TLS can negotiate dynamic WEP-keys which
> EAP/MD5 simply *can't*. FreeRADIUS now begins to support this dynamic
> WEP keys using Henrik and Lars patch and thanks to the recent changes.
> This patch is/will be integrated into 0.7.

Yes, I understand that TLS negotiates WEP keys dynamically based on the
certificate information shared between the AAA server and the
supplicant.  I don't think I explained myself well enough in my first
reply, as my terminology is lacking due to the fact that I am not
extreamely familiar with all of this quite yet.  Basically, I was
wondering if the EAP/MD5 combination does infact encrypt the data
between the client and AP, period.  Correct me if I am wrong, but this
method encrypts a preshared WEP key shared between the supplicant and
the AP using MD5?  That same MD5 challenge is used to encrypt the data
between the AP and the supplicant?  If yes, I know it's not as secure as
negotiated dynamic keys but none the less, it's encryption of some sort,
which is better than just a wep key and unencrypted data between
supplicant and AP.

Your explaination below is a little over my head, so please forgive me
if I've asked a question or made an assumption which is contrary to the
explaination you gave below in the first place :)

> 
> Explanation;
> 
> Every EAP mechanism being part of 802.1x basically defines the
> authentication scheme. So, neither of both has a direct regard to the
> air encryption, i.e. the RFCs of both specification do not talk about
> the used WEP encryption and WEP keys. The problem generally 
> is that from
> the point of view of the supplicant (user) the 
> EAP-communication ends at
> the server during the WEP encryption takes place between the 
> supplicant
> and the AP. According to the idea of 802.1x the AP will not 
> participate
> at the whichever EAP conversation. (It even couldn't in the 
> case of TLS,
> that's the sense of TLS). So, what to do?
> 
> However, EAP/TLS has a great advantage over EAP/MD5 because it
> negotiates the master keys (every TLS does). EAP/MD5 can't negotiate
> anything, it only verifies the identity with a rather weak method,
> cryptoanalitically spoken and compared to the PK-based TLS.
> 
> Anyway, some negotiated key material is available at the supplicant
> (user) and the authentication server after the TLS exchange. This
> material could be used to derive some other keys, which could then be
> sent to the AP (radius-client) using some (new) attributes. The
> supplicant already has the necessary key material since it 
> participated
> at the TLS exchange.
> 
> This advantage is used by MS and Cicso in their MPPE definition for
> WLANs (s. e.g. RFCs 3078 and 3079 and cisco documentation at
> www.cisco.com). The server adds the MPPE-Send-Key and MPPE-Recv-Key
> attributes, puts in these the derived key material and sends 
> them to the
> client (AP). AP has to understand these attributes. It then 
> derives the
> unicast and broadcast keys, encrypts them with the negotiated key and
> sends those in the EAPOL-Key message to the user. The user 
> uses its own
> key to decrypt the received EAPOL keys. These keys are setup 
> as dynamic
> WEP keys on the user side and at by the AP. An exact 
> description of the
> exchange is being prepared for as far as I know. Ask Raghu or 
> Henrik in
> case of doubt (or look in the RFCs :-))
> 
> The patch provided by Henrik and Lars adds the MPPE attributes to the
> server Accept message. There are people reporting that it works. In my
> case, with a Cicso AP340 I currently fail to activate it correctly:
> setting up dynamic WEP results in a communication breakdown. I'm
> currently on it, I hope the guys will help me ;-)
> 
> 
> That's it.
> 
> Artur
> 
> 
> 
> -- 
> Artur Hecker
> artur[at]hecker.info
> 
> - 
> List info/subscribe/unsubscribe? See 
> http://www.freeradius.org/list/users.html
> 



- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Reply via email to