Re: The crazy encryption madmans codebook
- From: "Joseph Ashwood" <ashwood@xxxxxxx>
- Date: Fri, 02 Mar 2007 08:32:14 GMT
<jt64@xxxxxxxx> wrote in message
news:1172820109.932982.192810@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On 1 Mar, 17:18, "Joseph Ashwood" <ashw...@xxxxxxx> wrote:
<j...@xxxxxxxx> wrote in message
news:1172793500.867394.322540@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
So your concept is to move the ancient concept of a codebook into a
database? It is ECB with encryption replaced by a lookup.
The overall security is as strong as what I will call the chaining mode
within and between the messages. By this I mean that is "orcdriffle"
always
translates to "telephone" the construct will leak and quickly fall apart.
No Joe i meant more like this
Actually you seem to have failed to understand what I was saying.
Consider 3 functions:
lp(string) looks up the given phrase in a database and returns a unique but
constant integer identifier for the word
li(integer) looks up the given integer identifier and returns a unique but
constant phrase associated with it
c(...) takes an integer as an input, computer a prp function on that input
and returns the value, both input and output are constrained to the possible
return values of li. Such that c' exists where x=c'(c(x)) for all valid x
This system of 3 functions is more powerful than your database lookup
system. There are 3 possible conditions:
c() takes only an integer (c(i))
c() takes an integer and a permutation identifier (key) (c(i,k))
c() takes an integer, a permutation identifier, and some state information
that it outputs for later input (c(i,k,v))
lp, li, c(i) combined are provably at least as powerful as your system as
they can express any system in your design. It is also trivally provable
that c(i) is less powerful than c(i,k)
lp,li,c(i,k) are trivially provable to be a cipher in ECB mode. As such they
fail in all the same ways as ECB with one major, and devastating exception,
each output represents a phrase and as such break accuracy can be quantified
more easily.
lp,li,c(i,k,v) represents the modification you proposed where k has length
0. c(i,k,v) has security equal to c(f(i,v),k) and that the security lies in
f(i,v) and c(i,k), seperating these out your multiplier has no entropic
input and as such for security purposes reduces to an identity, so the
security reduces to c(i,k) for which you have a 0 length k, so it reduces to
c(i) which is horrifying insecure.
As a result your entire construct, as it stands is extremely insecure. Also
it will be necessary to have a database with a prime number of entries
simply in order to make your proposal as difficult as possible to break
(still trivial though), so your proposed modification (in the other message)
actually lowers the security further because it is trivially provable that k
and k+1 cannot be prime for k>2.
So your entire system results in equivalence to a codebook, and your
modification to improve it actually weakens it.
Joe
.
- Follow-Ups:
- Re: The crazy encryption madmans codebook
- From: jt64
- Re: The crazy encryption madmans codebook
- From: jt64
- Re: The crazy encryption madmans codebook
- References:
- The crazy encryption madmans codebook
- From: jt64
- Re: The crazy encryption madmans codebook
- From: Joseph Ashwood
- Re: The crazy encryption madmans codebook
- From: jt64
- The crazy encryption madmans codebook
- Prev by Date: Re: disc erasure
- Next by Date: Re: disc erasure
- Previous by thread: Re: The crazy encryption madmans codebook
- Next by thread: Re: The crazy encryption madmans codebook
- Index(es):
Relevant Pages
|
