[Cryptography] A software for combining text files to obtain high quality pseudo-random sequences in practice

mok-kong shen mok-kong.shen at t-online.de
Tue Jul 11 03:58:57 EDT 2017


Am 11.07.2017 um 01:00 schrieb Sandy Harris:
> mok-kong shen <mok-kong.shen at t-online.de> wrote:
>
>> There are plenty of other schemes for obtaining high quality pseudo-random
>> sequences in practice, e.g. AES in counter mode. However our scheme seems to
>> be much simpler both in the underlying logic (understandability) and in
>> implementation and is thus a viable alternative ...
> I do not think it is viable, because it is not random; it depends
> entirely on the input text. One article puts total size of the
> web at 1.2 petabytes
> http://www.sciencefocus.com/qa/how-many-terabytes-data-are-internet
>
> Ignoring the facts that much of that is graphics, sound & video,
> and another bunch is in languages other than English.
> so your scheme might not work, and there's a whole lot
> of duplication ...
>
> That's 2^50 bytes so there are only 2^50 places to start
> your hash or other algorithm. If you use local text only,
> you might get 2^45 bytes (32 terabytes) or some such
> on a big RAID array with current technology.
>
> This is not nearly enough for any serious application,
> & you still need a 2^45 or 2^50-bit seed from a true
> RNG to choose a random starting point or it does
> not work at all.
>
> You are far better off using AES-CTR which has
> far lower overheads than your data access
> would require, and can take a larger key, 128
> bits to initialise the counter then another 128,
> 192 or 256 for the actual key.
>
>
>
> You can choose a stopping
> point too, and maybe use some local non-web text as
> well so it looks like you might get up to 2^64 or a bit
> more possible different outputs.
I have only tested the scheme for normal English texts. It's almost sure 
that it would
not have the same performance for special English texts or for some 
other languages,
not to say for the other kinds of information media. On the other hand, 
if it's good
enough for normal English, then, as I argued, it's a viable alternative 
one could use/need
under circumstances. (Note "under circumstances". and that I myself 
mentioned AES
in counter mode. ) Having workable alternatives (or Plan B) is IMHO 
advantageous in
general. Analogy: plenty of car owners also have bicycles which might 
become valuable
under circumstances. Certainly, I rely on Maurer's test and it is almost 
sure that it
wouldn't pass some more sophisticated statistical tests, but then this 
applies also to
plenty of other pseudo-random number generation schemes in practice, if 
I don't err.

M. K. Shen


More information about the cryptography mailing list