Re: OK What do you need? was:Re: Please test this encryption

From: Michael Brown (see_at_signature.below)
Date: 08/20/03 

Date: Wed, 20 Aug 2003 16:12:58 +1200

"LinuxGuy" <ilikelinux2001@yahoo.com> wrote in message
news:4b8dbf06.0308191019.63fc4868@posting.google.com...
> People claimed that a single sample was not enough for anyone to crack
> an encryption algorithm. I strongly disagree. For starters even
> strong encryption can be broken with brute force methods.

This is exactly why people are calling you a novice/idiot/etc. It's a
vacuous statement because *any* crypto (assuming you have more ciphertext
than key material and some idea of what the plaintext will look like) can be
broken by brute force. It's totally irrelevant to the discussion of whether
a short/single sample is sufficient to figure out an algorithm. It's easy to
show logically that a single ciphertext string is insufficient to determine
what the algorithm is: consider a normal block cypher. The encryption
algorithm is merely a bijective mapping from Z[2^N] numbers back to the same
set, ie: a permutation of the numbers 0,1,...,2^N. So given any plaintext, a
mapping can be found that generates the ciphertext.

Maybe an example would help ... suppose your ciphertext was
0101101010110001 = 23217 in decimal (the start of one of your messages)
If f(x)=x then the plaintext x = 23217 would work fine.
If f(x)=(modular inverse of x in Z[2^16+1]) then x = 63482 would be a valid
plaintext.
Any plaintext you give me, I can come up with a corresponding permutation
that will generate the ciphertext. So without the plaintext and/or the
algorithm, the ciphertext cannot be used to determine either of them.

Forutunately, you seem to have realised this to some extent with a
plaintext/ciphertext pair on your web page.

> Secondly
> weak encryption methods can also be broken simply by searching for
> patterns.

Sort of. Horrifically weak encryption can be broken as such. So your
algorithm isn't horifically weak. It still could be so weak I wouldn't even
want to encrypt my shopping list with it though.

> Another poster did a frequency analysis of the text, that should help
> those searching for patterns.
>
> Now tell me what else you would need? Anything other than the actual
> algorithm. Do you need a longer sample? How long? More samples? How
> many? Hell I can actually give you the key, without the proper
> application it wont work. And for those that are interested I have
> already posted another longer message (the thread is entitled "Is this
> long enough"?
> Now you have two messages,and a statistical analysis. If I am so
> incompetent, such a novice, this should be cracked in short order.

Again, I point to the fact that my terrible cipher could stand up to the
same scrutiny. 512kb of zeros would be a far better test (post it on your
webpage, copy 512kb from /dev/nul or whatever, I'd advise zipping it if
you're going to be using binary ascii).

One thing you don't explain on your page is how to actually read it. Is each
"..." thing a single block, or should all the strings under one heading be
appended into one great big stream?

One thing I do note is that all except two of the lines on your web page
start with "01". This is NOT a good sign.

> However I suspect I will just see another flurry of self righteous
> posts telling me that I am a novice, and idiot, read the FAQ,
> etc.....just crack it and you will have proven that its all
> true...until then you are just blathering.

I know you're going to call me a blatherer for this, but you don't seem to
get what the objections are. I would quite happily escrow $5000 into a bank
account for a year, post an identical challenge to yours using one of my
horribly weak encryption algorithms (undisclosed of course, I'd be ashamed
if anyone actually saw it), and I'd be 100% confident that I would keep my
cash for the year. This is precicely WHY challenges like yours are
ignored/flamed: they prove absolutely nothing about the security of the
algorithm, but people such as yourself feel they do. And I definately regard
myself as a crypto novice (I understand how linear/differential analysis
works on stuff like DES, but I doubt I'd be able to come up with a new
attack myself). 

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more :)
Add michael@ to emboss.co.nz - My inbox is always open

Relevant Pages

  • Re: New Encryption Idea
    ... performing the 5 reads necessary in the example algorithm results in a delay ... Panama at 400MB/sec, or RC4 at about 90MB/sec, or AES in CTR mode at ... and the speed failings of your design become very clear. ... > Manansala Encryption and Authentication System ...
    (sci.crypt)
  • Meganets "unbreakable" cryptography? Im skeptical.
    ... Meganet makes such grandiose claims that I can't help but ... There's plenty of coverage on secret encryption algorithms ... encryption algorithm that was granted U.S. Patent ... Labor has bought into this "snake oil" and without a doubt ...
    (sci.crypt)
  • Re: How good an encryption algorithm is this?
    ... >> know both a plaintext and a ciphertext for some message ... This is pretty much what happens with any encryption algorithm that uses ... XOR alone. ...
    (microsoft.public.dotnet.languages.csharp)
  • Re: How good an encryption algorithm is this?
    ... >> know both a plaintext and a ciphertext for some message ... This is pretty much what happens with any encryption algorithm that uses ... XOR alone. ...
    (microsoft.public.vc.language)
  • Re: Enigma machine strenght using a computer
    ... cipher with rotors that spin backwards, forwards, stop and start based ... on other rotors, change their values according to the phase of the moon ... encryption less error-prone and they were a cost effective way to get ... case) than not since a wider base uses the same algorithm. ...
    (sci.crypt)