Fangliang (Leon, ICSL) schrieb am 23.01.2015 um 08:54:
> The current Key Localization Algorithm have obvious vulnerability. We never 
> see such cryptographic algorithm defect in other widely used protocol. Hacker 
> communities have mentioned the current key localization method as a loophole, 
> and plan to exploit the vulnerability or publish it to CVE.

To me, your statement seems exaggerating the issue: The problem is simply that 
passwords A...A consisting of repetitions
of an expression A result in the same key as passwords with less repetitions. 
However, the entropy gained by repetitions
is quite low, and if a user chooses the password "bertbert", she cannot expect 
it much more difficult to guess than if
it was "bert". Thus, the fact that the attacker succeeds as soon as he guesses 
"bert" is just a gradual amplification of
an issue that exists outside the cryptographic method. And as others have 
pointed out, it is common sense that choosing
such passwords A...A is poor practice.

I don't mean say that the issue is completely irrelevant, but just it is not as 
severe as your post seems to imply. And
given the difficulties involved with a potential change, I doubt that it is 
worth.

That said, I suggest to include a warning into the security considerations of 
draft-hmac-sha-2-usm-snmp pointing to this
issue and recommending (generally) to use sufficiently strong passwords. RFC 
3414 already contains such a caveat but
repeating it in the new specification might help to draw attention to it.
-- 
Johannes

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to