Re: Random number generation

From: Tom St Denis (tomstdenis_at_iahu.ca)
Date: 06/27/03 

Date: Fri, 27 Jun 2003 16:44:59 GMT

Douglas Wikström wrote:
>>>I am just curious :-). Is there a rigorous security analysis of this
>>>construction.
>>
>>It's CTR mode. You're new here right?
>
>
> You are right, I misread the Simons post (my appologies to Simon :-). No
> Tom. It has been a while though. You have not changed.

What you mean my attitude? Go <expletive deleted> yourself you lame
<expletive deleted> piece of donkey <expletive deleted> mother
<expletive deleted> <expletive deleted> of donkey <expletive deleted>.

>>Essentially CTR mode is as secure as the underlying cipher is to known
>>and chosen plaintext attacks.
>
>
> What you are saying is not true in general. There are pathological
> counter examples. You need other properties from the underlying system.

Actually what I am saying is perfectly true and the prove is trivial.

If you can distinguish CTR mode from random with L ct/pt pairs then you
can break the cipher with L ct/pt pairs. Therefore if the cipher is
secure to N ct/pt pairs then CTR is secure to N ct/pt pairs.

If you mean "ooh ooh bitflip attack" then pay the <expletive deleted>
attention to the numerous <expletive deleted>'ing people who <expletive
deleted>'ing suggest to <expletive deleted>'ing use a <expletive
deleted>'ing MAC!

Tom 

--
LTC: http://libtomcrypt.org
LTM: http://math.libtomcrypt.org
The Book: http://book.libtomcrypt.org

Relevant Pages

  • Re: An idea against convention
    ... CTR mode is agnostic to your character ... >> encoding. ... This is ONLY meant for CTR mode. ... Tom ...
    (sci.crypt)
  • Re: A basic cryptanalysis question
    ... >>So provided your message length is a multiple of the block size I don't ... I'd use CTR mode in the first place. ... Nyah nyah! ... Tom ...
    (sci.crypt)
  • Re: Testing DES, 3DES and AES
    ... > Any particular reason you support all those bit resolutions for ... > CFB/OFB? ... Personally I'd say just go with CTR mode unless you're going ... Hi Tom, ...
    (sci.crypt)
  • Re: Data Compression Before or After Encryption ?
    ... that's not CTR mode. ... I don't think it would be insecure but ... Oh yeah, right. ... Listen to Tom. ...
    (sci.crypt)