[A83] Re: 83p os

[Prev][Next][Index][Thread]

[A83] Re: 83p os

Applications use rabin signature (hence the name rabsig for the old sign
tool) in combination with the md5.  Key files contain the modulus (p*q) and
then below that are the primes p and q which are congruent to 3 mod 4 (check
your own files the 2nd and 3rd line's 3 character will always be 3, 7, B, or
F because these are congruent to 3 mod 4) according to the requirement of
rabin signature scheme.  These large integers are in the format (two hex
bytes for size of number),(two hex bytes for LSB of number),(two hex bytes
for 2nd LSB of number),...,(two hex bytes for MSB of number).
Operating systems use the RSA algorithm in combination with the md5
signature.
For both these cases the modulus (p*q) comes out to be a 512 bit number.  To
sign your own os, the 512 bit number used for verifying an os would need to
be factored (or md5 would need to be broken which is also computationally
infeasible).

brandon sterner

----- Original Message -----
From: "David Phillips" <david@acz.org>
To: <assembly-83@lists.ticalc.org>
Sent: Monday, April 14, 2003 11:59 PM
Subject: [A83] Re: 83p os


> Jeff writes:
> > I agree, yet am confused at the reasoning behind factoring a 512 bit
> > number ;)
>
> Programs must be signed using RSA.  To sign your own code, you would need
> the key.  The key used on the calculator is 512 bits.  Thus, to sign your
> own code, you need to factor a 512 bit number.
>
> It is possible to factor a 512 bit number, but it would not be
economically
> feasible for you to do so.  It has been done in the past:
>
> http://www.rsasecurity.com/company/news/releases/pr.asp?doc_id=509
> http://citeseer.nj.nec.com/431185.html
> http://www.ercim.org/publication/Ercim_News/enw39/512.html
>
> There is an old BYTE article from October of 1995 which stated that 512
bit
> keys could be broken with then current technology:
>
> http://www.byte.com/art/9510/sec7/art5.htm
>
> Dan Bernstein has some very interesting ideas for special purpose circuits
> for factoring using number field sieve.  I would not be surprised if the
> next factoring breakthrough comes from him and his work:
>
> http://cr.yp.to/nfscircuit.html
> http://cr.yp.to/nfscircuit/rsa.html
>
> RSA Security has a new factoring challenge for factoring even larger
> numbers:
>
> http://www.rsasecurity.com/rsalabs/challenges/factoring/
>
> --
> David Phillips <david@acz.org>
> http://david.acz.org/
>
>
>
References: