Re: [Khufu] Pre-Computing the S-Boxes and obtaining aux. keys

From: Nikola Skoric (nick-news_at_net4u.hr)
Date: 10/31/04 

Date: Sun, 31 Oct 2004 13:39:14 +0100

Dana Sat, 30 Oct 2004 18:44:13 +0200
Mok-Kong Shen (mok-kong.shen@t-online.de) kaze...
> Nikola Skoric wrote:
> >
> > I'm trying to implement Khufu algoritm for my cryptography class. I
> > started with absolutely no information on my hands and achieved in
> > finding a paper (http://groups.google.com/groups?q=khufu+pre-
> > computing+s-boxes&hl=en&lr=&selm=7982%40hoptoad.uucp&rnum=2) written by
> > Ralph C. Merkle, the author of the algorithm, describing the
> > algorithm... the only problem is that's not enough to implement this
> > algorithm without making some wild assumptions.
>
> I am ignorant, but, according to HAC, the algorithm is patented.

Pardon me for my ignorance <g> but, what would HAC be?

> I don't know whether that would eventually cause you problems.

Well, if it caused problems, it would couse problems to my teacher who
put Khufu on the list of algoritmhs offered for implementation... so
that's fine :-)

> On the other hand, HAC says 'The majority of the patent consists
> of C-code listings specifying the ciphers'. So it seems that
> you might get quite some useful informations from the patent
> document.

Now, do you know where could I start asking for patent document?

> As to the generation of its S-boxes, which you asked, I have
> a 'guess':

Exactly what I'm trying to circumvent :-)

> The user chooses a PRNG and generates with the key
> as seed an array of 256 32-bit values. This would allow an
> 8-bit value to map to a 32-bit value, realizing the 8-32
> S-box. Of course, an 'essential' task that remains is
> appropriately finding one or a class of good PRNGs for that
> purpose.

Well, that for sure is really close to (if not exactly) the way Khufu
works, but I can't be sure because the word "seed" is not used to
connect the user supplied key and the stream of preudo-random numbers in
the original article... to be precise, there is no connection at all,
and that gap between the stream of preudo-random numbers and and the
user supplied key is my main problem... and if I connected that two, I'd
have the complete algorithm on my hands. 

-- 
"Stara boljka se leci starim lekom...
Dabome vinom, ta nebi valjda mlekom?"

Relevant Pages

  • Re: [Khufu] Pre-Computing the S-Boxes and obtaining aux. keys
    ... would you possibly know is there an online version of "Applied ... *might* be a full description of Khufu... ... "This is a quick hack to implement/test the Khufu encryption method ... algorithm in this code are hacks... ...
    (sci.crypt)
  • Re: [Khufu] Pre-Computing the S-Boxes and obtaining aux. keys
    ... > I'm trying to implement Khufu algoritm for my cryptography class. ... > Ralph C. Merkle, the author of the algorithm, describing the ... HAC says 'The majority of the patent consists ... The user chooses a PRNG and generates with the key ...
    (sci.crypt)
  • algoritam for pseudorandomisation
    ... I've been analizing Khufu block cipher algorithm and have come across an ... isn't detailed in the paper describing the Khufu algorithm. ... Knuth: "Seminumerical Algorithms", Vol II, Addison-Wesley Publishing ...
    (comp.programming)
  • Re: algoritam for pseudorandomisation
    ... > isn't detailed in the paper describing the Khufu algorithm. ... > algorithm for pseudorandomisation of an array of numbers, ... > description in the Knuth's book, and in the description of the Khufu ...
    (comp.programming)
  • Re: Here are some of my ideas for Fedora 8 and Fedora 9
    ... can the holder of the patent on an algorithm who has already been paid for a license to use on the device I have prevent me from using a different version of that same algorithm? ... Napster or the movie companies have gone after the people stealing movies ...
    (Fedora)