trying to reach one time key pad
category: code [glöplog]
heya guyse , i was just wondering , would it be close enough ( secure ) to one time key pad , if i tried implementing following on a id chip
1. i store some 355MB of random 0xFF numb res on the chip
2. i use mersenne twister ( predictable indices ) to choose i.e 16 ( or some random count ) strip from the key value ( which is stored somewhere on the server side too )
3. then do simple xor with the value i want to transfer securely
and all - if i'm not mistaken - is - i have to take care that the random index ( seed value ) - is shared between client and server
sounds too simple - but i simply can't see any flaws either
1. i store some 355MB of random 0xFF numb res on the chip
2. i use mersenne twister ( predictable indices ) to choose i.e 16 ( or some random count ) strip from the key value ( which is stored somewhere on the server side too )
3. then do simple xor with the value i want to transfer securely
and all - if i'm not mistaken - is - i have to take care that the random index ( seed value ) - is shared between client and server
sounds too simple - but i simply can't see any flaws either
as long as the 355M is used only once, then yes, that's a one time key pad.. =)
For any reasonable high amount of data no, it would not.
Basically, "Anyone can invent a security system that he himself cannot break."
Your system, will leak parts of your many-time-pad random buffer when randomly chosen strips begin to overlap. It's quite easy to calculate the probability of leaking any given percentage of your pad for any given amount of transferred data.
Basically, "Anyone can invent a security system that he himself cannot break."
Your system, will leak parts of your many-time-pad random buffer when randomly chosen strips begin to overlap. It's quite easy to calculate the probability of leaking any given percentage of your pad for any given amount of transferred data.
hmm and who knows - maybe fsqrt ( or someone else from his/her family ) - hooks up on the idea too.
- take osx into use
- start the app by choosing a random index, in the otherwise random executable, where the actual code is ( or do it 100 times if want to )
- put 1 or two easter eggs in the code ( using opengl ) to show how easy it's to take number 1 spot on the all time favs list if someone just wants to
- put .ogg player somewhere in the executable - or - separate player - if needed
- let the demo compo entry grow as big as allowed ( no source code included )
- take osx into use
- start the app by choosing a random index, in the otherwise random executable, where the actual code is ( or do it 100 times if want to )
- put 1 or two easter eggs in the code ( using opengl ) to show how easy it's to take number 1 spot on the all time favs list if someone just wants to
- put .ogg player somewhere in the executable - or - separate player - if needed
- let the demo compo entry grow as big as allowed ( no source code included )
also , if faqrt does't still hesitate , maybe this stupid piece of code from loosely 2005 finally works on all the browsers http://www.nbl.fi/~nbl2496/htm/looda.svg.html