Message-ID: <200605121049.40734.lucien.fransman@irc2.com>
Date: Fri May 12 19:55:34 2006
From: lucien.fransman at irc2.com (Lucien Fransman)
Subject: How secure is software X?

On Friday 12 May 2006 05:20, Blue Boar wrote:
Hello,
> Do you want just public results of standardized blackbox testing?
> Something similar to the ICSA firewall certification?  (Though, I assume
> you want actual public results.)
That would be ideal. properly anonimized ofcourse. It would be nice to have a 
list of applications that pass the "Litchfield" criteria. And software 
vendors would have a new marketing tool :) (Passed the  "Litchfield" 
criteria, so its better than product $foo, who failed 80% of the test)
>
> Would you include source review?  The Sardonix project tried to do that.
Well, ideally, that would be the case. But it would give an unfair advantage 
to software that has its source available. I would think its a nice extra, 
but not a requirement.  
>
> Who does the testing, and who pays for the time and equipment to do
> that?  Do all products get re-tested every time a new version of the
> product suite is released?  Do the test suites have to be free?  Do they
> re-test for every release of the victim software?
Well, the tests get done anyway. Why not bundle the results. The time spent at 
customer $bar is billed time anyway. using the (anonimized) results depends 
on NDA contracts and the like, but shouldn't pose a risk for that customer. 
And it would shorten the testing time in the future anyway. Because the 
pentester knows what the results of others are, and he only has to verify 
them.

>
> Don't people like yourself derive some benefit from having some portion
> of your assessment work stay proprietary?  If I'm trying to enhance the
> test suite with some new fuzzing, and I find a sexy bug, don't the
> incentives tend to lean towards me selling the bug to iDefense and
> hiding my fuzzer in the meantime?
I often wondered about this. An assessment is only as good as the assesser. 
What is the use of a "i can break and exploit $foo application, and have 
shown this in my tests", if it is done by a private exploit? Again, i'm 
thinking from the position of a company hiring a pentester/assesser, not by 
the multitude of people trying to gain from exploiting a 0day.
  
It only shows that the application has a bug, that is known to you or your 
company. Will it benefit the company that is being tested? I am not so sure 
about this. What would a company do with this kind of information? Fix the 
bug? They can't because they dont have access to the source.  Will it entice 
the vendor to fix the vulnerability? No, as they dont know it exists.
In my opinion, using private exploits and private vulnerabilities as a 
pentester/assesser only spreads FUD, and nothing really constructive is being 
done.
All this acomplishes is that the company who has the biggest archive of 
unpublished exploits, PoC code and vulnerabilities has a bigger chance at 
presenting a "we can break into your system 100% of the time" graph during 
sales talks.  
>
> Don't we fairly quickly arrive at all products passing all the standard
> tests, and "passing" no longer means anything?
It means something. It either means that the fuzzers or the testing technique 
is out of date, or that the applications being tested that pass have at least 
put some thought into the whole security process. Its a nice way to create a 
baseline. The application passess the default tests. that means that the 
applications have a minimum level of security. 

>
> I like the idea, but I'm wondering why people would contribute.  I'm
> also wondering how it can it stay consumer-beneficial, and not end up
> being driven by product vendors.
For the same reason as why OWASP and OSSTMM are successes. They dont get a lot 
of airtime, but everybody in the field knows about them. Everybody agrees (to 
a certain point) that they are useful. What would happen if we didn't have 
initiatives like this? There would be no framework, no comparison between 
methodologies and no standards.

The product vendors play a role, but I see it as the task of the people 
creating the standards to avoid it being a vendor platform only. And software 
vendors are not the enemy.

Anyway, these are my thoughts on this. The default disclaimers apply (not 
nessecarily the view of my employer, yadayada) 

>
> 						BB
Enchanter_tim

Powered by blists - more mailing lists