hasfunctions need not encrypt their data?

From: Immanuel Scholz (einwegadresse_at_gmx.de)
Date: 02/11/04 

Date: 11 Feb 2004 06:55:07 -0800

Hi group,

I just visited my final security related lesson at my university and
there are still some question left. ;-)

The most urgent:

Most papers, as example rfc2246 (TLS-Standard) or papers about HMAC's
refers to hash functions. The requirements for theses functions are
stated as collision free, means, "no two inputs can be found that
hashes to the same value".

But nowhere stated anything about "you may not reconstruct any
information of the original data from only looking at the hash value."

Is this a must requirement for each hash function? I do not think so.
I KNOW this is not a requirement for collision free, since you can
construct a collision free hash function that does not respect the
second sentence, if you have any collision free hash function by
simple concatenating pieces of the original message to the hash.

So why there is no written requirement that "hash functions must
encrypt their data they operate on"? Ok, most widely used Hashes like
MD5 or SHA are (hopefully?) secure about retrieving any original data
from the hash.

These features of hashes are used widely too - as example TLS uses
hashes as a pseudo random generator and it would be fatality if you
can easily reconstruct the seed from any part of the random string.

So why it is not explicitly a stated requirement for hash functions in
this scenarios?

Ciao, Imi.

PS: Can you please send carbon copy of answers to
"einwegadresse@gmx.de"? I do not recently check google-groups. ;)

PPS: Sorry for my bad english

Relevant Pages

  • Re: Calculating CRC32 for uploaded files
    ... >> There are other hash functions, but for files of this size you'd be ... crc32 hashes aren't unique while md5 hashes are. ... DeeDee, don't press that button! ...
    (comp.lang.php)
  • Re: Added hashes.
    ... Two hash functions are independent (with respect a ... convincing case that (md4 xor md5) is unbreakable? ... hashes, your definition of independence has to extend to those hashes. ... It can't rely on them being undistinguishable from random oracles. ...
    (sci.crypt)
  • Re: Hash functions strength
    ... Since hash functions are one-way, could, for example, the ... All Hashes are one way in the first sense. ... CRC is one way in the first sense but not in the second. ... ]Typical CRC functions are not one-way. ...
    (sci.crypt)
  • Re: Thou shalt have no other gods before the ANSI C standard
    ... >> height filled solely with research papers on string hashing, ... >> by the number of new papers being offered on the topic regularly, ... life as a test bed for hash functions, ... the broken "Reply" link at the bottom of the article. ...
    (sci.crypt)