On Feb 13, 2012, at 3:21 PM, Ted Lemon wrote:
On Feb 13, 2012, at 4:06 PM, Ben Campbell wrote:
Do
I infer correctly from your comment that the security properties of the mechanism don't really matter? That is, if the attacker we care about can't eavesdrop in the first place, does this really need to be an HMAC?
Hm, I thought about that a bit more after I wrote my response. The HMAC allows us to avoid sending the nonce in the clear in the DHCPFORCERENEW. I don't think this adds any value from a security perspective, actually, even though it feels more comfortable.
I suspect it was put in in the original version simply because of that—why send a secret over the wire when you don't have to? However, the original motivation for using the mechanism from RFC3315 was to avoid defining a new mechanism for a legacy protocol.
If we do need to change it, it's going to require a major rework of the document, and a lot of delay, so if it causes no harm, I think that's not worth doing.
I concur it's probably not worth changing at this point--but it does inform the "is MD5 good enough" discussion, which might make it worth mentioning in the security considerations.
I too would like to see the text I proposed added to the security considerations, so that we can be clear about what is being accomplished with this draft.
|
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf