On 19/03/2013 00:50, [email protected] wrote:
> So, my problem is to create a format-preserving injective function
> which is non-invertible (at least computationally).
>
> Since it's format preserving, it has to be a bijection, I'm guessing
> that basically boils down to a (computationally strong?) random
> permutation, for a domain of size =/= 2^n.

A hash with n-bit result, with input domain restricted to n-bit,
with n>=200 or so, is not a bijection, but is computationally
indistinguishable from what Iunderstand is your requirement.
Do I miss something?Or do you want it to be demonstrably
format-preserving/bijective(rather than computationally
indistinguishable from that)?

http://crypto.stackexchange.com would be another nice place to ask.

  Francois Grieu
_______________________________________________
cryptography mailing list
[email protected]
http://lists.randombit.net/mailman/listinfo/cryptography

Reply via email to