Re: C code of PEARL1, a block encryption algorithm emphasising simplicity


"Tom St Denis" <tom@xxxxxxx> wrote in message
news:4304b8f1-b407-4f79-88e4-

Well if it's a block cipher typically your plaintext can be longer
than a block? It was my vague understanding that he just gens a new
matrix for each block.

So if he encrypts a stream of identity blocks you get consecutive
blocks of PRNG data. If the PRNG is linear then you can break the
system.

If the matrix M is 4-by-4 then encrypting a stream over alphabet
{I1,I2,I3,I4} where (I1,I2,I3,I4) = I is the identity matrix, recovers
1/4 of M.

If the PRNG is secure [e.g. non-linear and non-correlated and large
enough search space] then why are you doing any of this nonsense
matrix work?

It is our view (yours and mine) that a PRNG is either secure or not. But
in Mok Kong Shen's view there are semi-secure (or 25%-secure,
20%-secure, 10%-secure etc) PRNGs that if 2N (or 4N, 5N, 10N etc) key
stream blocks is used to encrypt N plaintext blocks, it could be 100%
secure.

Regards,

Tran Ngoc Duong.


.

Relevant Pages

  • Re: Toaster to Generate Random Numbers
    ... >]A secure PRG is one for which if the input is fairly drawn, ... That is why people worry about a PRNG being ... The entropy of the output cannot be higher than the entropy of ...
    (comp.security.misc)
  • Re: Toaster to Generate Random Numbers
    ... >]A secure PRG is one for which if the input is fairly drawn, ... That is why people worry about a PRNG being ... The entropy of the output cannot be higher than the entropy of ...
    (sci.crypt)
  • Re: implementation of one-time-pad with Mersenne Twister PRNG
    ... generator is secure. ... The Mersenne Twister is not secure. ... PRNG, it is not sufficient to pass a set of commonly accepted statistical ... computers are quite good at it. ...
    (sci.crypt)
  • MT19337 for stream encryption?
    ... I am a novice at cryptography, so excuse my intrusion, but I'd like to ... Our application needs a very fast yet secure encryption method. ... We would, of course, forego using the PRNG seed generator for MT19337 ...
    (sci.crypt)
  • Re: MT19337 for stream encryption?
    ... opposed to using a weaker PRNG to do the job? ... hashing algorithm to make it truly secure? ... and simpler to achieve security. ... component in a cipher, even if it is a terrible cipher by itself. ...
    (sci.crypt)