Re: Security of Secret Algorithm encruption

William L. Bahn wrote:
> "Andrew Swallow" <am.swallow@btopenworld.com> wrote in message
> news:dghh12$4fi$1@nwrdmz02.dmz.ncs.ea.ibs-infra.bt.com...
>
>>William L. Bahn wrote:
>>
>>
>>>"Andrew Swallow" <am.swallow@btopenworld.com> wrote in
>
> message
>
>>>news:dgh60k$i7b$1@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com...
>>>
>>>
>>>>John E. Hadstate wrote:
>>>>
>>>>
>>>>
>>>>>"William L. Bahn" <william@toomuchspam.net> wrote in message
>>>>>news:-aGdnfWD9dFsVbbeRVn-3g@pcisys.net...
>>>>>
>>>>>
>>>>>
>>>>>>What I'm trying to get a feel for is how much harder
>>>>>>performing a
>>>>>>cryptanalysis is when the algorithms aren't known - or at
>>>>>>least
>>>>>>until they are known.
>>>>>
>>>>>
>>>>>Laying aside the technique of simply guessing a solution,
>>>>>the cryptanalyst must discover an algorithm that decrypts
>>>>>the ciphertext.
>>>>>
>>>>>I'm going to go out on a limb here and speculate that the
>>>>>minimum number of bits required to express this algorithm
>>>>>(which implicitly includes the "secret key") minus the
>>>>>amount of entropy in the "secret key" itself is the amount
>>>>>of entropy contributed by keeping the algorithm secret.
>>>>
>>>>The people making the code are probably beginners. All the
>>>>beginner's mistakes have known attacks. Simply go down the
>>>>list of attacks until one of them works.
>>>>
>>>
>>>
>>>In practice, I would certainly agree. But what if they didn't
>>>make all of the beginner's mistakes? Asked another way,
>
> perhaps,
>
>>>is how complicated does the "simplest" algorithm need to be
>
> while
>
>>>avoiding all of the "beginner's mistakes"?
>>
>>Simplest with no mistakes = AES or 3DES.
>>
>>The government does not pay extra for unnecessary complexity.
>
>
> And which we can also pretty safely assume that the attacker
> knows the algorithms for. The whole point of the thread is to get
> a feel for the degree of difficulty imposed by having to attack
> an algorithm when the algorithm is not known.

If you think about it I answered your question.
You need something the same complexity as AES.

Andrew Swallow
Received on Thu Sep 29 21:55:21 2005