Re:Theo de Raadt summarizes the OpenBSD IPSec "backdoor" sit...
by tsuraan on Tuesday December 21, 2010 @10:53PM
(#2030510)
What is the analogous mistake you'd make in CTR? The problem here was setting the IV to the last block of the previous stream, right? Would the analog to CTR be continuing to count in the same sequence? Also, I know this isn't the best forum for crypto discussion, but what is (especially for people who are almost totally ignorant but want to learn, like myself)?
[ Parent
| on HN ]
Re:Theo de Raadt summarizes the OpenBSD IPSec "backdoor" sit...
by tptacek on Tuesday December 21, 2010 @11:05PM
(#2030536)
The underlying ("moral", to coin a term) flaw isn't "using the last cipherblock of the last message as the IV for the next message; it's using predictable (or, worse, repeated) IVs.
An attacker who can get a target to repeat a CTR nonce can often coerce that target to repeat an entire keystream, which is devastating.
Re:Theo de Raadt summarizes the OpenBSD IPSec "backdoor" sit...
by yuhong on Saturday December 25, 2010 @12:44AM
(#2038587)
Yep, can be made with any stream cipher.
Re:Theo de Raadt summarizes the OpenBSD IPSec "backdoor" sit...
by tedunangst on Tuesday December 21, 2010 @11:25PM
(#2030563)
A good forum for discussion, or reading along, used to be sci.crypt with the spam filter turned up to 11+, but I haven't checked it out recently.