Re: [Full-Disclosure] MD5 hash cracking service

["Kurt Seifried" <listuser@xxxxxxxxxxxx>]

Minor correction: MD5 does suffer from potential collisions, that is two
different inputs can have the same MD5 output. Since the majority of systems
only store the MD5 and nothing else (like say what the first letter was, or
a SHA hash) if you find an input (not necessarily the "correct" one) that
results in the same MD5 hash the system will accept it as legitimate since
the hash values match. The chances of such a collision occuring, against
arbitrary data such as "this_is_my_password" are rare, especially when you
consider the limitations on passwords (length, ascii characters, etc.). The
beauty of the rainbow tables is once you've "cracked" a password (i.e. taken
a value, MD5'ed it, stored it in the DB) your only further requirement is
storage (which is dirt cheap now) and search costs (which is pretty cheap
since you have it all nicely done up in tables).

Kurt Seifried, kurt@xxxxxxxxxxxx
A15B BEE5 B391 B9AD B0EF
AEB0 AD63 0B4E AD56 E574
http://seifried.org/security/


_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html