Re: [PEN-TEST] Oracle USER$ password hashes (Summary)

[Stefan Aeschbacher <stefan () AESCHBACHER COM>]

Dragos Ruiu wrote:
[snip]
> Ok, I have to ask....
> Why has no-one disassembled oracle's hash program itself?
This most certainly will be the easier way to get the algorithm
than criptanalyzing some cleartex/ciphertext pairs. Still the
discussion led to some results, e.g. we found out, that a
birthday attack could be possible, this means, it is pretty
probable to find a (wrong) password which still hashes to
the same value as the original password as we have something
like 36^60 input and 2^64 output-values. Knowing the
structure of the algorithm used can help disassembling the
code regarding the form of input, output, salts used,....
As far as I'm concerned, I don't have access to an Oracle
nor have I the time to do it.

> The above effort seems like trying to reverse engineer a paper shredder
> by analyzing shredded paper instead of taking the device apart.
True, but some shredders in history have been analyzed this way without
even knowing whether they used fire, acid or mechanics to destroy the
paper ;) (see The Cryptonomicon by Donaldson or the books by D.Kahn for
some interesting reading on the subject)

Stefan