Re: Public disclosure of discovered vulnerabilities

brian_camillia_at_hotmail.com
Date: 05/15/05


Date: 15 May 2005 11:48:46 -0700

Well... that's a really difficult question without any perfect answer
--- two groups of people whose interests are in conflict but believing
they are the best!

But frankly, designing or implementing secure systems is subtle. If a
system is claimed to be provably secure, it's probably not! There are
always vulnerabilities around the corner when you declare your system
is secure. How could one gurantee he have already considered all kinds
of attacks or adversary capabilities when designing a system. Even
with provable security, what happens if the attacks come from
unanticipated adversary capabilities? Perhaps everyone of us should
keep in mind that the best promise one could give is a system will not
be broken in the forseeable future. When RSA was designed, who knew
the attacker would have the capabilities to launch side-channel
attacks, to measure the timing of computation, etc.. Computer
processors were still in their infant stage. Technology is advancing,
new attacks will definitely come as a consequence. Just blaming the
designer not having considered thorougly is not responsible, of course,
this does not include the careless ones. On the other hand, designers
should be more careful and open to feedback, not just close their minds
to think they have got the most innovative solution....

One question that bothered me for quite a while is what the people
working on cryptanalysis look for. What are their incentive? What are
their goals? Would people admire any destructive acts? I think most
of the hackers or attackers works for their ego, to show the world how
smart they are in showing others' mistakes. That's good! Without
them, all security partitioners would have lost their jobs. These
harsh wording are, of course, not for the researchers with a righteous
goal in locating a system's vulnerabilities ahead of the bad guys.
Frankly, we need the collaboration of both designer and hacker to build
a more secure system. When you start fire, I guess it would always be
a better question to ask what the response of average people would be
when I publish the weakness of a system they are using. Will they care
about how you can achieve such an attack? What they concern is "Well,
is it still secure enough for use? Any fix?". Here I would have my
humble suggestion to the people working on cryptanalysis: if you find a
new attack, it would always be good that you find a fix before you
spread out your incredible work. People tend to admire people with
constructive suggestions and work, and hate those bringing destruction.
 If you just want to show how smart you are in breaking others'
designs, there is no need to defend here --- a messenger is still
someone who brings troubles to all common people. If you have guts,
why don't you research on the vulnerabilities in the CIA, FBI or even
military networks and systems, and then publish your findings widely?
For sure, you will be arrested within hours for risking the safety of
the Great United States and charged for having the Weapons of Mass
Destruction. This is the common ground for the US government to accuse
other innocents!

Nicol So wrote:
> D. J. Bernstein wrote:
> >
> > Your shoot-the-messenger attitude reduces the punishment, and
therefore
> > reduces the incentive. You're seeing punishment as a bad thing
because
> > you're ignoring the massive long-term benefits of the incentive.
>
> Please stop mischaracterizing my position as "shooting the
messenger". I
> advocated no such thing.
>
> --
> Nicol So
> Disclaimer: Views expressed here are casual comments and should
> not be relied upon as the basis for decisions of consequence.



Relevant Pages

  • Re: My little something...
    ... There is no reason to belive it is LESS secure. ... The problem is unknown attacks are unknown ... I cant find any, can you? ... Chances that BOTH AES and WHATEVER are broken are SMALLER than chances ...
    (sci.crypt)
  • Re: My little something...
    ... There is no reason to belive it is LESS secure. ... The problem is unknown attacks are unknown ... I cant find any, can you? ... Chances that BOTH AES and WHATEVER are broken are SMALLER than chances ...
    (sci.crypt)
  • Re: More on AE / AEAD
    ... on the designerbeing /really/ good at noticing attacks. ... blockcipher wasn't secure. ... Having clear ideas of what `security' means for my finished system, ... using reductions from breaking my ...
    (sci.crypt)
  • Re: Is encrypting twice much more secure?
    ... The underlying DES ... DESX is still secure today. ... attacks work so well at establishing key bits from the signal [that is ...
    (sci.crypt)
  • Re: Illegal to do research on cryptography?
    ... > of] a standard algorithm would take longer than designing a secure ...
    (sci.crypt)