1
$\begingroup$

Whether the receiver can be sure that the message is sent by the source only and not spoofed by someone else?

$\endgroup$
1
  • $\begingroup$ There are 3 answers. Could you indicate if any answers your selection and accept the one that does? Otherwise please indicate what is missing... $\endgroup$
    – Maarten Bodewes
    Mar 12 '18 at 15:29
1
$\begingroup$

Does WPA(2) solve the problem of non-repudiation?

Whether the receiver can be sure that the message is sent by the source only and not spoofed by someone else?

You appear to be a bit confused about what non-repudiation actually means.

What it means that the receiver could prove to a third party that the sender (and not anyone else) sent the message. In this case, that is not true; the receiver could also generate the message.

On the other hand, in WPA(2), that's not what the receiver cares about. What the receiver does know is that someone with the WPA2 keys generated the message, and that (unless one of the two sides leaked it) those two sides are the sender and the receiver (and so if the receiver knows he didn't generate the message), the sender must have.

Hence, it proves what the receiver cares about (even if it doesn't meet the definition of "non-repudiation").

BTW: WPA uses a Message Authentication Code known as Michael; it was designed (in a hurry) specifically for WPA; it was designed under the constraint that it had to be implementable on every product that implemented WEP (hence doing something like HMAC, which would have been preferable, wasn't an option). Neils Fergusson did (IMHO) an excellent job in putting it together within the constraints he was given. However, I do believe some results have been found against it since; it'd be wiser to migrate to WPA2 (now that hardware that can do that is now available)

$\endgroup$
0
1
$\begingroup$

No it doesn't.

Message Authentication Codes (MACs, which in networking some may call MIC in order to differenciate from Media Access Control) are symmetric-key primitives that only provides authenticity and integrity of the message exchange. Non-repudiation requires public-key cryptography.

$\endgroup$
0
$\begingroup$

It depends, but usually the answer is no. The MAC is only used to provide integrity and authenticity of the message exchange.

Usually non-repudiation is something that is performed on documents rather than transport protocols. You could possibly prove that data was send through a transport protocol by a certain sender. This of course requires the sender to be authenticated within the protocol (e.g. client authentication for TLS). However, generally the packets of the transport protocol are discarded after the data has been received. If the MAC would be lost after receiving the data, the data cannot be proven to be signed by a specific entity anymore.

Certificate and keys used for authentication may also have different properties when it comes to real time authentication. For instance a certificate may have a lower key strength, validity time and may not be verified as strict as a certificate used for non-repudiation. Furthermore, a certificate / key use for authentication will not have the non-repudiation key usage set. Furthermore, an authentication key may also have less protection; smart cards, for instance, usually only require one PIN entry to use the authentication key multiple times. This would however only come to effect when WPA is used with EAP protocol for authentication.

More importantly, even if the data can be proven to be signed from a specific key, the sender may not consciously have signed that data. You could think of a JavaScript script that deliberately sends data through a WPA channel with client authentication. In other words: even if you can prove that you received an otherwise unsigned document agreeing to sell my house, I won't have sold you my house. I can repute it: I may never have consciously send it to you and I certainly did not sign it with the intention to sell the house.

If you're stupid enough to use a unique password, leak that password to a third party and then establish a Wireless connection using WPA and then leak the data and keying material, somebody may be able to prove that you have send that data. That would not be non-repudiation as any party receiving the keys might just as well have signed the data themselves, but it may break the closely related notion of deniability.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.