On Thu, 17 Feb 2005, Craig J. Lindstrom wrote:

> Security is all about probabilities anyway. There are an infinite number of
> collisions for every hash value.  This does not make hashing useless.  It
> appears that SHA-1 is not as "strong" as was thought. Big deal, no one ever
> believed it was that strong anyway.  So go to SHA-256 or SHA-512.  They are
> both better(as far as we know) but neither will prevent collisions.  The
> point is, at what level are you comfortable with your security?

Sigh.  It IS a big deal.  People DID think that SHA-1 was secure until quite
recently.  That's why it was a NIST standard.  It has nothing to do with
preventing collisions from existing.  It has everything to do with making them
too hard to find.

The point is often what level of security you're comfortable with, but in this
case it's "are you comfortable using a broken system", and since you're on the
UUG and not the Windows user group list, I presume the answer is no.


> Throughout history every time a locksmith builds a better lock, a thief
> eventually acquires the skills to pick it.  No surprise here.  We just move
> on, and build a better lock.

It is unsurprising when a new algorithm is broken.  It is much more surprising
when a hash function endorsed by NIST and beaten on for many years falls.  
That's what's happened here.  3DES and RSA have stood for 20 years now, and
hopefully we'll find a new hash function that will last along with them.


>  My point is that the world as we know it has not ended.  Sure we need to be
> aware of things, and take appropriate action.  But don't quit locking the
> door just because someone can break in, most people can't or won't.

>  On another point, if you are concerned about security why would anyone use
> a self-made pgp key?  It is not trustworthy by nature.  Sort of like making
> your own social security card.

>  Security is a many layered beast. We just make it hard enough that it is
> improbable to get at whatever we are protecting.  It is never impossible.  
> Also you have to consider cost, you should not spend more on security than
> the value of that which you are securing.  Granted the "value" of many
> things is hard to ascertain.  Remember layers upon layers.

Thanks for teaching us this important lesson about security.  Being as the
topic is the apparent discovery of collisions in the SHA-1 hash, though,
perhaps you and the other pundits could have checked your facts a little
better on that topic before proclaiming that the world is a safe place and
that we need not pay any mind to what the actual security researchers are
saying.

To answer your question about PGP, it's every bit as secure as SSL or SSH in
the scenario where you personally verify the fingerprints of the keys you
receive.  For people you don't know personally, it uses a decentralized web of
trust, rather than the trusted-CA model used by SSL.  (SSH doesn't bother with
either, and expects you to verify each fingerprint manually).  The 'pedia 
knows all:
http://en.wikipedia.org/wiki/Web_of_trust

You can even see a picture of the keysigning party we did in CS465:
http://en.wikipedia.org/wiki/Keysigning_party

                                                -J


--------------------
BYU Unix Users Group 
http://uug.byu.edu/ 

The opinions expressed in this message are the responsibility of their
author.  They are not endorsed by BYU, the BYU CS Department or BYU-UUG. 
___________________________________________________________________
List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list

Reply via email to