[lug] Interesting sum "weakness"

[Tkil]

more information on the MD5 issues:

http://www.math.fu-berlin.de/~guckes/md5/

   gives references to the collision work (scroll down to the section
   entitled "MD5 - Security").

http://www.vpnc.org/ietf-ipsec/97.ipsec/msg01855.html

   a view from the ipsec side of things.  incidental, and no obvious
   references that i can see, but it does help one to think of the
   situations in which a "ten hour collision construction" may or may
   not be important when using MD5.

http://www.uni-mainz.de/~pommeren/DSVorlesung/Material/MD5.Dobbertin

   words from the author himself.

http://the.wiretapped.net/security/textfiles/crypto-misc/standard-cryptographic-algorithm-naming.html

   has links to a PS version of the note from Hans Dobbertin as well
   as the note from RSA regarding his attacks.  Quoting:

      Given the surprising speed with which techniques on MD4 were
      extended to MD5 we feel that it is only prudent to draw a
      cautious conclusion and to expect that collisions for the entire
      hash function might soon be found.

   the link given is to:  

ftp://ftp.rsa.com/pub/pdfs/bulletn4.pdf

   a condensed note (only 6 pages, reasonably easy reading, lots of
   background) that explains current status of MD2, MD4, MD5 (all
   effectively broken, MD5 "least broken" and therefore possibly still
   useful for some purposes), and alternatives.  note that MD5 was
   released into the public domain (most likely because they wanted it
   integrated into a standard which would refuse proprietary or
   patented solutions), so RSA has no stake in keeping its weaknesses
   hidden.

enough web searching for now,
t.