Re: Unique number generation

paul_at_atom.sbrk.co.uk
Date: 04/29/05


Date: Fri, 29 Apr 2005 12:48:30 GMT

In article <d4t6cv$lsg$1@orkan.itea.ntnu.no>, Kristian Gjøsteen wrote:
><paul@atom.sbrk.co.uk> wrote:
>>Any ideas?
>
> Use a block cipher: value = E_k(sequence number).

Doh. That is a little too obvious... Ok, I also should'be said that the
output value needs to be constrained within a certain range, so was
thinking of a hash. However, that's not really possible without the risk
of collisions. Ok, looks like I need to constrain my inputs.

Cheers,

Paul



Relevant Pages

  • Re: Unique number generation
    ... > Doh. ... > thinking of a hash. ... looks like I need to constrain my inputs. ... length -- checking for collisions is not hard, ...
    (sci.crypt)
  • Re: Unique number generation
    ... >Doh. ... Try your luck and Luby-Rackoff with ... google. ... Kristian Gjøsteen ...
    (sci.crypt)