# Properties of Ideal Straight P-Boxes

What properties should an ideal straight P-Box exhibit?

-
Could you add a definition of "straight P-Box", for those of us who don't know it? – Paŭlo Ebermann Jan 7 '13 at 17:20

Straight P-Boxes should have only 1 input for each output and the same number of inputs as outputs.

Straight P-Boxes will also be invert-able.

-
I love the default dictionary! – ponsfonze Jan 5 '13 at 17:49
OK, removed that, but where is "ideal", shouldn't it be "Ideal P-Boxes will also be invert-able"? – Maarten Bodewes Jan 5 '13 at 18:13
OK, looked it up, and it should be Straight P-Boxes will also be invert-able. Still missing the word "ideal " in the answer though. – Maarten Bodewes Jan 5 '13 at 20:13
You probably also want to add that the P-Boxes need to maximize diffusion across inputs and outputs, to make linear and differential cryptanalysis more difficult. Optimal diffusion = ideal P-Box, I suppose. – Thomas Jan 6 '13 at 2:46
I will try and revise my answer to have the complete information if it is missing anything. – ponsfonze Jan 7 '13 at 18:41

This is not an answerable question. The requirements for a P-box (whatever that is) depend upon the cipher it is embedded in; that cipher imposes some requirements on the P-box. Therefore, you cannot do P-box design in the abstract, without knowing what the requirements are. There is no one, single answer that is valid for all ciphers.

P.S. Where did you hear the term "P-box"? As far as I know, this is not an accepted or widely used term, let alone "straight P-box".

-
The question is about "straight P-boxes". I have no idea what these are, but it doesn't sound same as "S-boxes". – Paŭlo Ebermann Jan 7 '13 at 17:19
@PaŭloEbermann, I see that I mis-read the question; I thought it said S-box. A search on "straight P-box" finds that Wikipedia says a straight P-box is a permutation (reordering) of $n$ bits. It's just a transposition (re-ordering) of the bits. Still, "ideal" cannot be defined without reference to the block cipher in which it be used. – D.W. Jan 7 '13 at 20:05