Re: MD5 collisions


"Mike Amling" <nospam@xxxxxxxxxx> wrote in message
news:eaqvis$qbl@xxxxxxxxxxxxxxxxxxxxxxxxxx
..
..

For any bit sequences str_a and str_b of any length, if str_a and
str_b have the same MD5 hash, then it is certain that str_a || x ||
str_c collides with str_b || y || str_c, where x is the padding used by
MD5 for str_a and y is the padding used by MD5 for str_b, and str_c is
any sequence of bits.
It is also certain that if str_a and str_b have the same length and
end in the same 64 bytes then str_a || str_c collides with str_b ||
str_c for any bit sequence str_c.


Thanks for that additional info, Mike. I take it that || is the
concatenation operator. (Please correct me if that is not the case.)

The string 'the quick brown fox jumps over the lazy dog' hashes to
77add1d5f41223d5582fca736a5cb335.

What are the current (rough) estimates on how long it would take to find a
different string that hashes to the same value - on, say, a 2GHz home
computer. (If it's still a matter of brute-forcing then I can work that out
for myself .. but I'm thinking there might now be better methods.)

If it's a feasible task, is there any software around that already achieves
that task ? (Again, the question pretty much answers itself if brute force
is still the only method. I imagine that fear of use for nefarious purposes
might make such software difficult to locate ... if it even exists.)

Thanks guys.

Cheers,
Rob


.

Relevant Pages
Quantcast