[ossec-dev] [Bug 52] Proposal to use a hash more secure

[ossec-bugzilla@xxxxxxxxxxxxxxxxxxxxxxxxxxx]

http://www.ossec.net/bugs/show_bug.cgi?id=52





------- Comment #5 from jonas.esp@xxxxxxxxxxxxxx  2007-08-16 07:27 -------
(In reply to comment #3)
> (In reply to comment #2)
> > Well, and would not it be better to use one only algorithm --but secure-- which
> > will have a better performance that using two weak algorithms simultaneously?
> > 
> 
> Weak? sha1 is not considered weak. And today, md5 is not considered weak
> either.
> Is it not better to use 2 different algorithms, rather than rely on the one
> (Tiger)? 
> 
> I think we are getting ahead of ourselves as md5 is still secure today. And we
> are creating "defense in depth" by using another algorithm as well. 
> 
> It is not as if we have performance issues by using 2 hashing algorithms.
> 
It only is necessary to looking for anything as "md5 hash weak" (the same with
sha1) to know if those algorithms are secure. A hash algorithm isn't secure
since that *have known hash collision weaknesses*.

It is also is very interesting looking for information related to Bruce
Schneier about those algorithms. And I'm supposed that he knows what is saying.
;)

I'm going to writing the conclusions since this paper (about MD5):
http://packetstorm.linuxsecurity.com/papers/cryptography/md5_someday.pdf

"The point is not that MD5 has collapsed. It hasn’t. The point is that
there’s a very clear trend regarding the security level of MD5, and it
isn’t good. It is now undeniable that the selection of MD5 matters – the
constraint that deployed implementations of the one-way hash primitive be
functionally identical has been broken. The failures detected are not merely
algorithmic or theoretical, rather new capabilities above and beyond what the
primitive specifies are made available by the selection of MD5. It is not
expected that this paper will cause a precipitous decline in the use of MD5;
that will probably occur when a means of silently introducing single-bit errors
in arbitrary (rather than chosen) MD5 payloads is discovered.

But in the security community, we tend to complain about the ”phase change”
nature of our systems that suddenly collapse from secure to insecure on the
discovery of a ”zero day” exploit. The phase change for MD5 isn’t here
yet, but it will come, someday. Nobody should be surprised when that day
arrives."


-- 
Configure bugmail: http://www.ossec.net/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.